This is an automated email from the ASF dual-hosted git repository.
github-actions[bot] pushed a commit to branch gh-pages
in repository https://gitbox.apache.org/repos/asf/iceberg-rust.git
The following commit(s) were added to refs/heads/gh-pages by this push:
new 05abf07e3 deploy: de19a3986fbdf399f09f71cd8c08e2573630c80f
05abf07e3 is described below
commit 05abf07e378a494bcdf74d738234e4dc774de8fd
Author: github-merge-queue[bot]
<github-merge-queue[bot]@users.noreply.github.com>
AuthorDate: Thu Oct 1 08:09:55 2026 +0000
deploy: de19a3986fbdf399f09f71cd8c08e2573630c80f
---
api/src/iceberg/arrow/delete_filter.rs.html | 811 ++++++++++++++--------------
1 file changed, 408 insertions(+), 403 deletions(-)
diff --git a/api/src/iceberg/arrow/delete_filter.rs.html
b/api/src/iceberg/arrow/delete_filter.rs.html
index ab5f2758c..6be202bdd 100644
--- a/api/src/iceberg/arrow/delete_filter.rs.html
+++ b/api/src/iceberg/arrow/delete_filter.rs.html
@@ -168,417 +168,422 @@
<a href=#168 id=168 data-nosnippet>168</a> <span
class="kw-2">&</span><span class="self">self</span>,
<a href=#169 id=169 data-nosnippet>169</a> file_path: <span
class="kw-2">&</span>str,
<a href=#170 id=170 data-nosnippet>170</a> ) -> <span
class="prelude-ty">Option</span><Predicate> {
-<a href=#171 id=171 data-nosnippet>171</a> <span class="kw">let
</span>notifier = {
-<a href=#172 id=172 data-nosnippet>172</a> <span class="kw">match
</span><span
class="self">self</span>.state.read().unwrap().equality_deletes.get(file_path) {
-<a href=#173 id=173 data-nosnippet>173</a> <span
class="prelude-val">None </span>=> <span class="kw">return </span><span
class="prelude-val">None</span>,
-<a href=#174 id=174 data-nosnippet>174</a> <span
class="prelude-val">Some</span>(EqDelState::Loading(notifier)) =>
notifier.clone(),
-<a href=#175 id=175 data-nosnippet>175</a> <span
class="prelude-val">Some</span>(EqDelState::Loaded(predicate)) => {
-<a href=#176 id=176 data-nosnippet>176</a> <span
class="kw">return </span><span
class="prelude-val">Some</span>(predicate.clone());
-<a href=#177 id=177 data-nosnippet>177</a> }
-<a href=#178 id=178 data-nosnippet>178</a> }
-<a href=#179 id=179 data-nosnippet>179</a> };
-<a href=#180 id=180 data-nosnippet>180</a>
-<a href=#181 id=181 data-nosnippet>181</a> notifier.notified().<span
class="kw">await</span>;
-<a href=#182 id=182 data-nosnippet>182</a>
-<a href=#183 id=183 data-nosnippet>183</a> <span class="kw">match
</span><span
class="self">self</span>.state.read().unwrap().equality_deletes.get(file_path) {
-<a href=#184 id=184 data-nosnippet>184</a> <span
class="prelude-val">Some</span>(EqDelState::Loaded(predicate)) => <span
class="prelude-val">Some</span>(predicate.clone()),
-<a href=#185 id=185 data-nosnippet>185</a> <span class="kw">_
</span>=> <span class="macro">unreachable!</span>(<span
class="string">"Cannot be any other state than loaded"</span>),
-<a href=#186 id=186 data-nosnippet>186</a> }
-<a href=#187 id=187 data-nosnippet>187</a> }
-<a href=#188 id=188 data-nosnippet>188</a>
-<a href=#189 id=189 data-nosnippet>189</a> <span class="doccomment">///
Builds eq delete predicate for the provided task.
-<a href=#190 id=190 data-nosnippet>190</a> </span><span
class="kw">pub</span>(<span class="kw">crate</span>) <span class="kw">async fn
</span>build_equality_delete_predicate(
-<a href=#191 id=191 data-nosnippet>191</a> <span
class="kw-2">&</span><span class="self">self</span>,
-<a href=#192 id=192 data-nosnippet>192</a> file_scan_task: <span
class="kw-2">&</span>FileScanTask,
-<a href=#193 id=193 data-nosnippet>193</a> ) -> <span
class="prelude-ty">Result</span><<span
class="prelude-ty">Option</span><BoundPredicate>> {
-<a href=#194 id=194 data-nosnippet>194</a> <span class="comment">// *
Filter the task's deletes into just the Equality deletes
-<a href=#195 id=195 data-nosnippet>195</a> // * Retrieve the unbound
predicate for each from self.state.equality_deletes
-<a href=#196 id=196 data-nosnippet>196</a> // * Logical-AND them all
together to get a single combined `Predicate`
-<a href=#197 id=197 data-nosnippet>197</a> // * Bind the predicate to
the task's schema to get a `BoundPredicate`
-<a href=#198 id=198 data-nosnippet>198</a>
-<a href=#199 id=199 data-nosnippet>199</a> </span><span class="kw">let
</span><span class="kw-2">mut </span>combined_predicate = AlwaysTrue;
-<a href=#200 id=200 data-nosnippet>200</a> <span class="kw">for
</span>delete <span class="kw">in </span>file_scan_task.deletes() {
-<a href=#201 id=201 data-nosnippet>201</a> <span class="kw">if
</span>!is_equality_delete(delete) {
-<a href=#202 id=202 data-nosnippet>202</a> <span
class="kw">continue</span>;
-<a href=#203 id=203 data-nosnippet>203</a> }
-<a href=#204 id=204 data-nosnippet>204</a>
-<a href=#205 id=205 data-nosnippet>205</a> <span class="kw">let
</span><span class="prelude-val">Some</span>(predicate) = <span
class="self">self
-<a href=#206 id=206 data-nosnippet>206</a>
</span>.get_equality_delete_predicate_for_delete_file_path(<span
class="kw-2">&</span>delete.file_path)
-<a href=#207 id=207 data-nosnippet>207</a> .<span
class="kw">await
-<a href=#208 id=208 data-nosnippet>208</a> else </span>{
-<a href=#209 id=209 data-nosnippet>209</a> <span
class="kw">return </span><span class="prelude-val">Err</span>(Error::new(
-<a href=#210 id=210 data-nosnippet>210</a>
ErrorKind::Unexpected,
-<a href=#211 id=211 data-nosnippet>211</a> <span
class="macro">format!</span>(
-<a href=#212 id=212 data-nosnippet>212</a> <span
class="string">"Missing predicate for equality delete file '{}'"</span>,
-<a href=#213 id=213 data-nosnippet>213</a>
delete.file_path
-<a href=#214 id=214 data-nosnippet>214</a> ),
-<a href=#215 id=215 data-nosnippet>215</a> ));
-<a href=#216 id=216 data-nosnippet>216</a> };
-<a href=#217 id=217 data-nosnippet>217</a>
-<a href=#218 id=218 data-nosnippet>218</a> combined_predicate =
combined_predicate.and(predicate);
-<a href=#219 id=219 data-nosnippet>219</a> }
-<a href=#220 id=220 data-nosnippet>220</a>
-<a href=#221 id=221 data-nosnippet>221</a> <span class="kw">if
</span>combined_predicate == AlwaysTrue {
-<a href=#222 id=222 data-nosnippet>222</a> <span class="kw">return
</span><span class="prelude-val">Ok</span>(<span
class="prelude-val">None</span>);
-<a href=#223 id=223 data-nosnippet>223</a> }
-<a href=#224 id=224 data-nosnippet>224</a>
-<a href=#225 id=225 data-nosnippet>225</a> <span class="kw">let
</span>bound_predicate = combined_predicate
-<a href=#226 id=226 data-nosnippet>226</a>
.bind(file_scan_task.schema_ref(), file_scan_task.case_sensitive())<span
class="question-mark">?</span>;
-<a href=#227 id=227 data-nosnippet>227</a> <span
class="prelude-val">Ok</span>(<span
class="prelude-val">Some</span>(bound_predicate))
-<a href=#228 id=228 data-nosnippet>228</a> }
+<a href=#171 id=171 data-nosnippet>171</a> <span class="comment">//
Build the `Notified` while holding the read lock. `notified_owned()` records
tokio's
+<a href=#172 id=172 data-nosnippet>172</a> // `notify_waiters_calls`
counter at construction and completes on first poll if that
+<a href=#173 id=173 data-nosnippet>173</a> // counter has since
advanced. Reading the counter under the lock guarantees it is taken
+<a href=#174 id=174 data-nosnippet>174</a> // before
`insert_equality_delete` can advance it via `notify_waiters()`, so the
+<a href=#175 id=175 data-nosnippet>175</a> // notification is never
missed even though we `.await` after releasing the lock.
+<a href=#176 id=176 data-nosnippet>176</a> </span><span class="kw">let
</span>notified = {
+<a href=#177 id=177 data-nosnippet>177</a> <span class="kw">match
</span><span
class="self">self</span>.state.read().unwrap().equality_deletes.get(file_path) {
+<a href=#178 id=178 data-nosnippet>178</a> <span
class="prelude-val">None </span>=> <span class="kw">return </span><span
class="prelude-val">None</span>,
+<a href=#179 id=179 data-nosnippet>179</a> <span
class="prelude-val">Some</span>(EqDelState::Loading(notifier)) =>
notifier.clone().notified_owned(),
+<a href=#180 id=180 data-nosnippet>180</a> <span
class="prelude-val">Some</span>(EqDelState::Loaded(predicate)) => {
+<a href=#181 id=181 data-nosnippet>181</a> <span
class="kw">return </span><span
class="prelude-val">Some</span>(predicate.clone());
+<a href=#182 id=182 data-nosnippet>182</a> }
+<a href=#183 id=183 data-nosnippet>183</a> }
+<a href=#184 id=184 data-nosnippet>184</a> };
+<a href=#185 id=185 data-nosnippet>185</a>
+<a href=#186 id=186 data-nosnippet>186</a> notified.<span
class="kw">await</span>;
+<a href=#187 id=187 data-nosnippet>187</a>
+<a href=#188 id=188 data-nosnippet>188</a> <span class="kw">match
</span><span
class="self">self</span>.state.read().unwrap().equality_deletes.get(file_path) {
+<a href=#189 id=189 data-nosnippet>189</a> <span
class="prelude-val">Some</span>(EqDelState::Loaded(predicate)) => <span
class="prelude-val">Some</span>(predicate.clone()),
+<a href=#190 id=190 data-nosnippet>190</a> <span class="kw">_
</span>=> <span class="macro">unreachable!</span>(<span
class="string">"Cannot be any other state than loaded"</span>),
+<a href=#191 id=191 data-nosnippet>191</a> }
+<a href=#192 id=192 data-nosnippet>192</a> }
+<a href=#193 id=193 data-nosnippet>193</a>
+<a href=#194 id=194 data-nosnippet>194</a> <span class="doccomment">///
Builds eq delete predicate for the provided task.
+<a href=#195 id=195 data-nosnippet>195</a> </span><span
class="kw">pub</span>(<span class="kw">crate</span>) <span class="kw">async fn
</span>build_equality_delete_predicate(
+<a href=#196 id=196 data-nosnippet>196</a> <span
class="kw-2">&</span><span class="self">self</span>,
+<a href=#197 id=197 data-nosnippet>197</a> file_scan_task: <span
class="kw-2">&</span>FileScanTask,
+<a href=#198 id=198 data-nosnippet>198</a> ) -> <span
class="prelude-ty">Result</span><<span
class="prelude-ty">Option</span><BoundPredicate>> {
+<a href=#199 id=199 data-nosnippet>199</a> <span class="comment">// *
Filter the task's deletes into just the Equality deletes
+<a href=#200 id=200 data-nosnippet>200</a> // * Retrieve the unbound
predicate for each from self.state.equality_deletes
+<a href=#201 id=201 data-nosnippet>201</a> // * Logical-AND them all
together to get a single combined `Predicate`
+<a href=#202 id=202 data-nosnippet>202</a> // * Bind the predicate to
the task's schema to get a `BoundPredicate`
+<a href=#203 id=203 data-nosnippet>203</a>
+<a href=#204 id=204 data-nosnippet>204</a> </span><span class="kw">let
</span><span class="kw-2">mut </span>combined_predicate = AlwaysTrue;
+<a href=#205 id=205 data-nosnippet>205</a> <span class="kw">for
</span>delete <span class="kw">in </span>file_scan_task.deletes() {
+<a href=#206 id=206 data-nosnippet>206</a> <span class="kw">if
</span>!is_equality_delete(delete) {
+<a href=#207 id=207 data-nosnippet>207</a> <span
class="kw">continue</span>;
+<a href=#208 id=208 data-nosnippet>208</a> }
+<a href=#209 id=209 data-nosnippet>209</a>
+<a href=#210 id=210 data-nosnippet>210</a> <span class="kw">let
</span><span class="prelude-val">Some</span>(predicate) = <span
class="self">self
+<a href=#211 id=211 data-nosnippet>211</a>
</span>.get_equality_delete_predicate_for_delete_file_path(<span
class="kw-2">&</span>delete.file_path)
+<a href=#212 id=212 data-nosnippet>212</a> .<span
class="kw">await
+<a href=#213 id=213 data-nosnippet>213</a> else </span>{
+<a href=#214 id=214 data-nosnippet>214</a> <span
class="kw">return </span><span class="prelude-val">Err</span>(Error::new(
+<a href=#215 id=215 data-nosnippet>215</a>
ErrorKind::Unexpected,
+<a href=#216 id=216 data-nosnippet>216</a> <span
class="macro">format!</span>(
+<a href=#217 id=217 data-nosnippet>217</a> <span
class="string">"Missing predicate for equality delete file '{}'"</span>,
+<a href=#218 id=218 data-nosnippet>218</a>
delete.file_path
+<a href=#219 id=219 data-nosnippet>219</a> ),
+<a href=#220 id=220 data-nosnippet>220</a> ));
+<a href=#221 id=221 data-nosnippet>221</a> };
+<a href=#222 id=222 data-nosnippet>222</a>
+<a href=#223 id=223 data-nosnippet>223</a> combined_predicate =
combined_predicate.and(predicate);
+<a href=#224 id=224 data-nosnippet>224</a> }
+<a href=#225 id=225 data-nosnippet>225</a>
+<a href=#226 id=226 data-nosnippet>226</a> <span class="kw">if
</span>combined_predicate == AlwaysTrue {
+<a href=#227 id=227 data-nosnippet>227</a> <span class="kw">return
</span><span class="prelude-val">Ok</span>(<span
class="prelude-val">None</span>);
+<a href=#228 id=228 data-nosnippet>228</a> }
<a href=#229 id=229 data-nosnippet>229</a>
-<a href=#230 id=230 data-nosnippet>230</a> <span
class="kw">pub</span>(<span class="kw">crate</span>) <span class="kw">fn
</span>upsert_delete_vector(
-<a href=#231 id=231 data-nosnippet>231</a> <span class="kw-2">&mut
</span><span class="self">self</span>,
-<a href=#232 id=232 data-nosnippet>232</a> data_file_path: String,
-<a href=#233 id=233 data-nosnippet>233</a> delete_vector: DeleteVector,
-<a href=#234 id=234 data-nosnippet>234</a> ) {
-<a href=#235 id=235 data-nosnippet>235</a> <span class="kw">let
</span><span class="kw-2">mut </span>state = <span
class="self">self</span>.state.write().unwrap();
-<a href=#236 id=236 data-nosnippet>236</a>
-<a href=#237 id=237 data-nosnippet>237</a> <span class="kw">let
</span><span class="prelude-val">Some</span>(entry) =
state.delete_vectors.get_mut(<span class="kw-2">&</span>data_file_path)
<span class="kw">else </span>{
-<a href=#238 id=238 data-nosnippet>238</a> state
-<a href=#239 id=239 data-nosnippet>239</a> .delete_vectors
-<a href=#240 id=240 data-nosnippet>240</a>
.insert(data_file_path, Arc::new(Mutex::new(delete_vector)));
-<a href=#241 id=241 data-nosnippet>241</a> <span
class="kw">return</span>;
-<a href=#242 id=242 data-nosnippet>242</a> };
-<a href=#243 id=243 data-nosnippet>243</a>
-<a href=#244 id=244 data-nosnippet>244</a> <span
class="kw-2">*</span>entry.lock().unwrap() |= delete_vector;
-<a href=#245 id=245 data-nosnippet>245</a> }
-<a href=#246 id=246 data-nosnippet>246</a>
-<a href=#247 id=247 data-nosnippet>247</a> <span
class="kw">pub</span>(<span class="kw">crate</span>) <span class="kw">fn
</span>insert_equality_delete(
-<a href=#248 id=248 data-nosnippet>248</a> <span
class="kw-2">&</span><span class="self">self</span>,
-<a href=#249 id=249 data-nosnippet>249</a> delete_file_path: <span
class="kw-2">&</span>str,
-<a href=#250 id=250 data-nosnippet>250</a> eq_del:
Receiver<Predicate>,
-<a href=#251 id=251 data-nosnippet>251</a> ) {
-<a href=#252 id=252 data-nosnippet>252</a> <span class="kw">let
</span>notify = Arc::new(Notify::new());
-<a href=#253 id=253 data-nosnippet>253</a> {
-<a href=#254 id=254 data-nosnippet>254</a> <span class="kw">let
</span><span class="kw-2">mut </span>state = <span
class="self">self</span>.state.write().unwrap();
-<a href=#255 id=255 data-nosnippet>255</a>
state.equality_deletes.insert(
-<a href=#256 id=256 data-nosnippet>256</a>
delete_file_path.to_string(),
-<a href=#257 id=257 data-nosnippet>257</a>
EqDelState::Loading(notify.clone()),
-<a href=#258 id=258 data-nosnippet>258</a> );
-<a href=#259 id=259 data-nosnippet>259</a> }
-<a href=#260 id=260 data-nosnippet>260</a>
-<a href=#261 id=261 data-nosnippet>261</a> <span class="kw">let
</span>state = <span class="self">self</span>.state.clone();
-<a href=#262 id=262 data-nosnippet>262</a> <span class="kw">let
</span>delete_file_path = delete_file_path.to_string();
-<a href=#263 id=263 data-nosnippet>263</a> <span
class="self">self</span>.runtime.cpu().spawn(<span class="kw">async move
</span>{
-<a href=#264 id=264 data-nosnippet>264</a> <span class="kw">let
</span>eq_del = eq_del.<span class="kw">await</span>.unwrap();
-<a href=#265 id=265 data-nosnippet>265</a> {
-<a href=#266 id=266 data-nosnippet>266</a> <span class="kw">let
</span><span class="kw-2">mut </span>state = state.write().unwrap();
-<a href=#267 id=267 data-nosnippet>267</a> state
-<a href=#268 id=268 data-nosnippet>268</a> .equality_deletes
-<a href=#269 id=269 data-nosnippet>269</a>
.insert(delete_file_path, EqDelState::Loaded(eq_del));
-<a href=#270 id=270 data-nosnippet>270</a> }
-<a href=#271 id=271 data-nosnippet>271</a> notify.notify_waiters();
-<a href=#272 id=272 data-nosnippet>272</a> });
-<a href=#273 id=273 data-nosnippet>273</a> }
-<a href=#274 id=274 data-nosnippet>274</a>}
-<a href=#275 id=275 data-nosnippet>275</a>
-<a href=#276 id=276 data-nosnippet>276</a><span class="kw">pub</span>(<span
class="kw">crate</span>) <span class="kw">fn </span>is_equality_delete(f: <span
class="kw-2">&</span>FileScanTaskDeleteFile) -> bool {
-<a href=#277 id=277 data-nosnippet>277</a> <span
class="macro">matches!</span>(f.file_type, DataContentType::EqualityDeletes)
-<a href=#278 id=278 data-nosnippet>278</a>}
-<a href=#279 id=279 data-nosnippet>279</a>
-<a href=#280 id=280 data-nosnippet>280</a><span class="attr">#[cfg(test)]
-<a href=#281 id=281 data-nosnippet>281</a></span><span
class="kw">pub</span>(<span class="kw">crate</span>) <span class="kw">mod
</span>tests {
-<a href=#282 id=282 data-nosnippet>282</a> <span class="kw">use
</span>std::fs::File;
-<a href=#283 id=283 data-nosnippet>283</a> <span class="kw">use
</span>std::path::Path;
-<a href=#284 id=284 data-nosnippet>284</a> <span class="kw">use
</span>std::sync::Arc;
-<a href=#285 id=285 data-nosnippet>285</a>
-<a href=#286 id=286 data-nosnippet>286</a> <span class="kw">use
</span>arrow_array::{Int64Array, RecordBatch, StringArray};
-<a href=#287 id=287 data-nosnippet>287</a> <span class="kw">use
</span>arrow_schema::Schema <span class="kw">as </span>ArrowSchema;
-<a href=#288 id=288 data-nosnippet>288</a> <span class="kw">use
</span>parquet::arrow::{ArrowWriter, PARQUET_FIELD_ID_META_KEY};
-<a href=#289 id=289 data-nosnippet>289</a> <span class="kw">use
</span>parquet::basic::Compression;
-<a href=#290 id=290 data-nosnippet>290</a> <span class="kw">use
</span>parquet::file::properties::WriterProperties;
-<a href=#291 id=291 data-nosnippet>291</a> <span class="kw">use
</span>tempfile::TempDir;
-<a href=#292 id=292 data-nosnippet>292</a>
-<a href=#293 id=293 data-nosnippet>293</a> <span class="kw">use
</span>super::<span class="kw-2">*</span>;
-<a href=#294 id=294 data-nosnippet>294</a> <span class="kw">use
</span><span
class="kw">crate</span>::arrow::caching_delete_file_loader::CachingDeleteFileLoader;
-<a href=#295 id=295 data-nosnippet>295</a> <span class="kw">use
</span><span class="kw">crate</span>::expr::Reference;
-<a href=#296 id=296 data-nosnippet>296</a> <span class="kw">use
</span><span class="kw">crate</span>::io::FileIO;
-<a href=#297 id=297 data-nosnippet>297</a> <span class="kw">use
</span><span class="kw">crate</span>::spec::{DataFileFormat, Datum,
NestedField, PrimitiveType, Schema, Type};
-<a href=#298 id=298 data-nosnippet>298</a>
-<a href=#299 id=299 data-nosnippet>299</a> <span class="kw">type
</span>ArrowSchemaRef = Arc<ArrowSchema>;
-<a href=#300 id=300 data-nosnippet>300</a>
-<a href=#301 id=301 data-nosnippet>301</a> <span class="kw">const
</span>FIELD_ID_POSITIONAL_DELETE_FILE_PATH: u64 = <span
class="number">2147483546</span>;
-<a href=#302 id=302 data-nosnippet>302</a> <span class="kw">const
</span>FIELD_ID_POSITIONAL_DELETE_POS: u64 = <span
class="number">2147483545</span>;
+<a href=#230 id=230 data-nosnippet>230</a> <span class="kw">let
</span>bound_predicate = combined_predicate
+<a href=#231 id=231 data-nosnippet>231</a>
.bind(file_scan_task.schema_ref(), file_scan_task.case_sensitive())<span
class="question-mark">?</span>;
+<a href=#232 id=232 data-nosnippet>232</a> <span
class="prelude-val">Ok</span>(<span
class="prelude-val">Some</span>(bound_predicate))
+<a href=#233 id=233 data-nosnippet>233</a> }
+<a href=#234 id=234 data-nosnippet>234</a>
+<a href=#235 id=235 data-nosnippet>235</a> <span
class="kw">pub</span>(<span class="kw">crate</span>) <span class="kw">fn
</span>upsert_delete_vector(
+<a href=#236 id=236 data-nosnippet>236</a> <span class="kw-2">&mut
</span><span class="self">self</span>,
+<a href=#237 id=237 data-nosnippet>237</a> data_file_path: String,
+<a href=#238 id=238 data-nosnippet>238</a> delete_vector: DeleteVector,
+<a href=#239 id=239 data-nosnippet>239</a> ) {
+<a href=#240 id=240 data-nosnippet>240</a> <span class="kw">let
</span><span class="kw-2">mut </span>state = <span
class="self">self</span>.state.write().unwrap();
+<a href=#241 id=241 data-nosnippet>241</a>
+<a href=#242 id=242 data-nosnippet>242</a> <span class="kw">let
</span><span class="prelude-val">Some</span>(entry) =
state.delete_vectors.get_mut(<span class="kw-2">&</span>data_file_path)
<span class="kw">else </span>{
+<a href=#243 id=243 data-nosnippet>243</a> state
+<a href=#244 id=244 data-nosnippet>244</a> .delete_vectors
+<a href=#245 id=245 data-nosnippet>245</a>
.insert(data_file_path, Arc::new(Mutex::new(delete_vector)));
+<a href=#246 id=246 data-nosnippet>246</a> <span
class="kw">return</span>;
+<a href=#247 id=247 data-nosnippet>247</a> };
+<a href=#248 id=248 data-nosnippet>248</a>
+<a href=#249 id=249 data-nosnippet>249</a> <span
class="kw-2">*</span>entry.lock().unwrap() |= delete_vector;
+<a href=#250 id=250 data-nosnippet>250</a> }
+<a href=#251 id=251 data-nosnippet>251</a>
+<a href=#252 id=252 data-nosnippet>252</a> <span
class="kw">pub</span>(<span class="kw">crate</span>) <span class="kw">fn
</span>insert_equality_delete(
+<a href=#253 id=253 data-nosnippet>253</a> <span
class="kw-2">&</span><span class="self">self</span>,
+<a href=#254 id=254 data-nosnippet>254</a> delete_file_path: <span
class="kw-2">&</span>str,
+<a href=#255 id=255 data-nosnippet>255</a> eq_del:
Receiver<Predicate>,
+<a href=#256 id=256 data-nosnippet>256</a> ) {
+<a href=#257 id=257 data-nosnippet>257</a> <span class="kw">let
</span>notify = Arc::new(Notify::new());
+<a href=#258 id=258 data-nosnippet>258</a> {
+<a href=#259 id=259 data-nosnippet>259</a> <span class="kw">let
</span><span class="kw-2">mut </span>state = <span
class="self">self</span>.state.write().unwrap();
+<a href=#260 id=260 data-nosnippet>260</a>
state.equality_deletes.insert(
+<a href=#261 id=261 data-nosnippet>261</a>
delete_file_path.to_string(),
+<a href=#262 id=262 data-nosnippet>262</a>
EqDelState::Loading(notify.clone()),
+<a href=#263 id=263 data-nosnippet>263</a> );
+<a href=#264 id=264 data-nosnippet>264</a> }
+<a href=#265 id=265 data-nosnippet>265</a>
+<a href=#266 id=266 data-nosnippet>266</a> <span class="kw">let
</span>state = <span class="self">self</span>.state.clone();
+<a href=#267 id=267 data-nosnippet>267</a> <span class="kw">let
</span>delete_file_path = delete_file_path.to_string();
+<a href=#268 id=268 data-nosnippet>268</a> <span
class="self">self</span>.runtime.cpu().spawn(<span class="kw">async move
</span>{
+<a href=#269 id=269 data-nosnippet>269</a> <span class="kw">let
</span>eq_del = eq_del.<span class="kw">await</span>.unwrap();
+<a href=#270 id=270 data-nosnippet>270</a> {
+<a href=#271 id=271 data-nosnippet>271</a> <span class="kw">let
</span><span class="kw-2">mut </span>state = state.write().unwrap();
+<a href=#272 id=272 data-nosnippet>272</a> state
+<a href=#273 id=273 data-nosnippet>273</a> .equality_deletes
+<a href=#274 id=274 data-nosnippet>274</a>
.insert(delete_file_path, EqDelState::Loaded(eq_del));
+<a href=#275 id=275 data-nosnippet>275</a> }
+<a href=#276 id=276 data-nosnippet>276</a> notify.notify_waiters();
+<a href=#277 id=277 data-nosnippet>277</a> });
+<a href=#278 id=278 data-nosnippet>278</a> }
+<a href=#279 id=279 data-nosnippet>279</a>}
+<a href=#280 id=280 data-nosnippet>280</a>
+<a href=#281 id=281 data-nosnippet>281</a><span class="kw">pub</span>(<span
class="kw">crate</span>) <span class="kw">fn </span>is_equality_delete(f: <span
class="kw-2">&</span>FileScanTaskDeleteFile) -> bool {
+<a href=#282 id=282 data-nosnippet>282</a> <span
class="macro">matches!</span>(f.file_type, DataContentType::EqualityDeletes)
+<a href=#283 id=283 data-nosnippet>283</a>}
+<a href=#284 id=284 data-nosnippet>284</a>
+<a href=#285 id=285 data-nosnippet>285</a><span class="attr">#[cfg(test)]
+<a href=#286 id=286 data-nosnippet>286</a></span><span
class="kw">pub</span>(<span class="kw">crate</span>) <span class="kw">mod
</span>tests {
+<a href=#287 id=287 data-nosnippet>287</a> <span class="kw">use
</span>std::fs::File;
+<a href=#288 id=288 data-nosnippet>288</a> <span class="kw">use
</span>std::path::Path;
+<a href=#289 id=289 data-nosnippet>289</a> <span class="kw">use
</span>std::sync::Arc;
+<a href=#290 id=290 data-nosnippet>290</a>
+<a href=#291 id=291 data-nosnippet>291</a> <span class="kw">use
</span>arrow_array::{Int64Array, RecordBatch, StringArray};
+<a href=#292 id=292 data-nosnippet>292</a> <span class="kw">use
</span>arrow_schema::Schema <span class="kw">as </span>ArrowSchema;
+<a href=#293 id=293 data-nosnippet>293</a> <span class="kw">use
</span>parquet::arrow::{ArrowWriter, PARQUET_FIELD_ID_META_KEY};
+<a href=#294 id=294 data-nosnippet>294</a> <span class="kw">use
</span>parquet::basic::Compression;
+<a href=#295 id=295 data-nosnippet>295</a> <span class="kw">use
</span>parquet::file::properties::WriterProperties;
+<a href=#296 id=296 data-nosnippet>296</a> <span class="kw">use
</span>tempfile::TempDir;
+<a href=#297 id=297 data-nosnippet>297</a>
+<a href=#298 id=298 data-nosnippet>298</a> <span class="kw">use
</span>super::<span class="kw-2">*</span>;
+<a href=#299 id=299 data-nosnippet>299</a> <span class="kw">use
</span><span
class="kw">crate</span>::arrow::caching_delete_file_loader::CachingDeleteFileLoader;
+<a href=#300 id=300 data-nosnippet>300</a> <span class="kw">use
</span><span class="kw">crate</span>::expr::Reference;
+<a href=#301 id=301 data-nosnippet>301</a> <span class="kw">use
</span><span class="kw">crate</span>::io::FileIO;
+<a href=#302 id=302 data-nosnippet>302</a> <span class="kw">use
</span><span class="kw">crate</span>::spec::{DataFileFormat, Datum,
NestedField, PrimitiveType, Schema, Type};
<a href=#303 id=303 data-nosnippet>303</a>
-<a href=#304 id=304 data-nosnippet>304</a> <span class="comment">//
Regression test for the positional-delete lost-wakeup hang.
-<a href=#305 id=305 data-nosnippet>305</a> //
-<a href=#306 id=306 data-nosnippet>306</a> // Drives the real API through
the losing interleaving: the loader fires
-<a href=#307 id=307 data-nosnippet>307</a> // `notify_waiters()` (via
`finish_pos_del_load`) *before* the waiter awaits the notifier
-<a href=#308 id=308 data-nosnippet>308</a> // handed back by `WaitFor`.
`notify_waiters()` stores no permit, so this only completes if
-<a href=#309 id=309 data-nosnippet>309</a> // the waiter's `Notified` was
created before the signal. Because `WaitFor` now carries an
-<a href=#310 id=310 data-nosnippet>310</a> // `OwnedNotified` created under
the lock in `try_start_pos_del_load`, it is; on the old
-<a href=#311 id=311 data-nosnippet>311</a> // `WaitFor(Arc<Notify>)`
contract the waiter created its `Notified` too late and hung.
-<a href=#312 id=312 data-nosnippet>312</a> </span><span
class="attr">#[tokio::test]
-<a href=#313 id=313 data-nosnippet>313</a> </span><span class="kw">async fn
</span>test_wait_for_completes_when_load_finishes_before_await() {
-<a href=#314 id=314 data-nosnippet>314</a> <span class="kw">let
</span>filter = DeleteFilter::new(Runtime::current());
-<a href=#315 id=315 data-nosnippet>315</a> <span class="kw">let
</span>path = <span class="string">"s3://bucket/pos-delete.parquet"</span>;
-<a href=#316 id=316 data-nosnippet>316</a>
-<a href=#317 id=317 data-nosnippet>317</a> <span
class="macro">assert!</span>(<span class="macro">matches!</span>(
-<a href=#318 id=318 data-nosnippet>318</a>
filter.try_start_pos_del_load(path),
-<a href=#319 id=319 data-nosnippet>319</a> PosDelLoadAction::Load
-<a href=#320 id=320 data-nosnippet>320</a> ));
+<a href=#304 id=304 data-nosnippet>304</a> <span class="kw">type
</span>ArrowSchemaRef = Arc<ArrowSchema>;
+<a href=#305 id=305 data-nosnippet>305</a>
+<a href=#306 id=306 data-nosnippet>306</a> <span class="kw">const
</span>FIELD_ID_POSITIONAL_DELETE_FILE_PATH: u64 = <span
class="number">2147483546</span>;
+<a href=#307 id=307 data-nosnippet>307</a> <span class="kw">const
</span>FIELD_ID_POSITIONAL_DELETE_POS: u64 = <span
class="number">2147483545</span>;
+<a href=#308 id=308 data-nosnippet>308</a>
+<a href=#309 id=309 data-nosnippet>309</a> <span class="comment">//
Regression test for the positional-delete lost-wakeup hang.
+<a href=#310 id=310 data-nosnippet>310</a> //
+<a href=#311 id=311 data-nosnippet>311</a> // Drives the real API through
the losing interleaving: the loader fires
+<a href=#312 id=312 data-nosnippet>312</a> // `notify_waiters()` (via
`finish_pos_del_load`) *before* the waiter awaits the notifier
+<a href=#313 id=313 data-nosnippet>313</a> // handed back by `WaitFor`.
`notify_waiters()` stores no permit, so this only completes if
+<a href=#314 id=314 data-nosnippet>314</a> // the waiter's `Notified` was
created before the signal. Because `WaitFor` now carries an
+<a href=#315 id=315 data-nosnippet>315</a> // `OwnedNotified` created under
the lock in `try_start_pos_del_load`, it is; on the old
+<a href=#316 id=316 data-nosnippet>316</a> // `WaitFor(Arc<Notify>)`
contract the waiter created its `Notified` too late and hung.
+<a href=#317 id=317 data-nosnippet>317</a> </span><span
class="attr">#[tokio::test]
+<a href=#318 id=318 data-nosnippet>318</a> </span><span class="kw">async fn
</span>test_wait_for_completes_when_load_finishes_before_await() {
+<a href=#319 id=319 data-nosnippet>319</a> <span class="kw">let
</span>filter = DeleteFilter::new(Runtime::current());
+<a href=#320 id=320 data-nosnippet>320</a> <span class="kw">let
</span>path = <span class="string">"s3://bucket/pos-delete.parquet"</span>;
<a href=#321 id=321 data-nosnippet>321</a>
-<a href=#322 id=322 data-nosnippet>322</a> <span class="kw">let
</span>PosDelLoadAction::WaitFor(notified) =
filter.try_start_pos_del_load(path) <span class="kw">else </span>{
-<a href=#323 id=323 data-nosnippet>323</a> <span
class="macro">panic!</span>(<span class="string">"expected WaitFor for an
in-progress load"</span>);
-<a href=#324 id=324 data-nosnippet>324</a> };
-<a href=#325 id=325 data-nosnippet>325</a>
-<a href=#326 id=326 data-nosnippet>326</a> <span class="comment">//
Loader completes and signals before the waiter awaits.
-<a href=#327 id=327 data-nosnippet>327</a>
</span>filter.finish_pos_del_load(path);
-<a href=#328 id=328 data-nosnippet>328</a>
-<a href=#329 id=329 data-nosnippet>329</a> <span class="kw">let
</span>waited = tokio::time::timeout(std::time::Duration::from_secs(<span
class="number">5</span>), notified).<span class="kw">await</span>;
-<a href=#330 id=330 data-nosnippet>330</a> <span
class="macro">assert!</span>(
-<a href=#331 id=331 data-nosnippet>331</a> waited.is_ok(),
-<a href=#332 id=332 data-nosnippet>332</a> <span
class="string">"WaitFor future must resolve after finish_pos_del_load"
-<a href=#333 id=333 data-nosnippet>333</a> </span>);
-<a href=#334 id=334 data-nosnippet>334</a> }
-<a href=#335 id=335 data-nosnippet>335</a>
-<a href=#336 id=336 data-nosnippet>336</a> <span class="attr">#[tokio::test]
-<a href=#337 id=337 data-nosnippet>337</a> </span><span class="kw">async fn
</span>test_delete_file_filter_load_deletes() {
-<a href=#338 id=338 data-nosnippet>338</a> <span class="kw">let
</span>tmp_dir = TempDir::new().unwrap();
-<a href=#339 id=339 data-nosnippet>339</a> <span class="kw">let
</span>table_location = tmp_dir.path();
-<a href=#340 id=340 data-nosnippet>340</a> <span class="kw">let
</span>file_io = FileIO::new_with_fs();
-<a href=#341 id=341 data-nosnippet>341</a>
-<a href=#342 id=342 data-nosnippet>342</a> <span class="kw">let
</span>delete_file_loader =
-<a href=#343 id=343 data-nosnippet>343</a>
CachingDeleteFileLoader::new(file_io.clone(), <span class="number">10</span>,
Runtime::current());
-<a href=#344 id=344 data-nosnippet>344</a>
-<a href=#345 id=345 data-nosnippet>345</a> <span class="kw">let
</span>file_scan_tasks = setup(table_location);
+<a href=#322 id=322 data-nosnippet>322</a> <span
class="macro">assert!</span>(<span class="macro">matches!</span>(
+<a href=#323 id=323 data-nosnippet>323</a>
filter.try_start_pos_del_load(path),
+<a href=#324 id=324 data-nosnippet>324</a> PosDelLoadAction::Load
+<a href=#325 id=325 data-nosnippet>325</a> ));
+<a href=#326 id=326 data-nosnippet>326</a>
+<a href=#327 id=327 data-nosnippet>327</a> <span class="kw">let
</span>PosDelLoadAction::WaitFor(notified) =
filter.try_start_pos_del_load(path) <span class="kw">else </span>{
+<a href=#328 id=328 data-nosnippet>328</a> <span
class="macro">panic!</span>(<span class="string">"expected WaitFor for an
in-progress load"</span>);
+<a href=#329 id=329 data-nosnippet>329</a> };
+<a href=#330 id=330 data-nosnippet>330</a>
+<a href=#331 id=331 data-nosnippet>331</a> <span class="comment">//
Loader completes and signals before the waiter awaits.
+<a href=#332 id=332 data-nosnippet>332</a>
</span>filter.finish_pos_del_load(path);
+<a href=#333 id=333 data-nosnippet>333</a>
+<a href=#334 id=334 data-nosnippet>334</a> <span class="kw">let
</span>waited = tokio::time::timeout(std::time::Duration::from_secs(<span
class="number">5</span>), notified).<span class="kw">await</span>;
+<a href=#335 id=335 data-nosnippet>335</a> <span
class="macro">assert!</span>(
+<a href=#336 id=336 data-nosnippet>336</a> waited.is_ok(),
+<a href=#337 id=337 data-nosnippet>337</a> <span
class="string">"WaitFor future must resolve after finish_pos_del_load"
+<a href=#338 id=338 data-nosnippet>338</a> </span>);
+<a href=#339 id=339 data-nosnippet>339</a> }
+<a href=#340 id=340 data-nosnippet>340</a>
+<a href=#341 id=341 data-nosnippet>341</a> <span class="attr">#[tokio::test]
+<a href=#342 id=342 data-nosnippet>342</a> </span><span class="kw">async fn
</span>test_delete_file_filter_load_deletes() {
+<a href=#343 id=343 data-nosnippet>343</a> <span class="kw">let
</span>tmp_dir = TempDir::new().unwrap();
+<a href=#344 id=344 data-nosnippet>344</a> <span class="kw">let
</span>table_location = tmp_dir.path();
+<a href=#345 id=345 data-nosnippet>345</a> <span class="kw">let
</span>file_io = FileIO::new_with_fs();
<a href=#346 id=346 data-nosnippet>346</a>
-<a href=#347 id=347 data-nosnippet>347</a> <span class="kw">let
</span>delete_filter = delete_file_loader
-<a href=#348 id=348 data-nosnippet>348</a> .load_deletes(
-<a href=#349 id=349 data-nosnippet>349</a>
file_scan_tasks[<span class="number">0</span>].deletes(),
-<a href=#350 id=350 data-nosnippet>350</a>
file_scan_tasks[<span class="number">0</span>].schema_ref(),
-<a href=#351 id=351 data-nosnippet>351</a> )
-<a href=#352 id=352 data-nosnippet>352</a> .<span class="kw">await
-<a href=#353 id=353 data-nosnippet>353</a> </span>.unwrap()
-<a href=#354 id=354 data-nosnippet>354</a> .unwrap();
-<a href=#355 id=355 data-nosnippet>355</a>
-<a href=#356 id=356 data-nosnippet>356</a> <span class="kw">let
</span>result = delete_filter
-<a href=#357 id=357 data-nosnippet>357</a> .get_delete_vector(<span
class="kw-2">&</span>file_scan_tasks[<span class="number">0</span>])
-<a href=#358 id=358 data-nosnippet>358</a> .unwrap();
-<a href=#359 id=359 data-nosnippet>359</a> <span
class="macro">assert_eq!</span>(result.lock().unwrap().len(), <span
class="number">12</span>); <span class="comment">// pos dels from pos del file
1 and 2
+<a href=#347 id=347 data-nosnippet>347</a> <span class="kw">let
</span>delete_file_loader =
+<a href=#348 id=348 data-nosnippet>348</a>
CachingDeleteFileLoader::new(file_io.clone(), <span class="number">10</span>,
Runtime::current());
+<a href=#349 id=349 data-nosnippet>349</a>
+<a href=#350 id=350 data-nosnippet>350</a> <span class="kw">let
</span>file_scan_tasks = setup(table_location);
+<a href=#351 id=351 data-nosnippet>351</a>
+<a href=#352 id=352 data-nosnippet>352</a> <span class="kw">let
</span>delete_filter = delete_file_loader
+<a href=#353 id=353 data-nosnippet>353</a> .load_deletes(
+<a href=#354 id=354 data-nosnippet>354</a>
file_scan_tasks[<span class="number">0</span>].deletes(),
+<a href=#355 id=355 data-nosnippet>355</a>
file_scan_tasks[<span class="number">0</span>].schema_ref(),
+<a href=#356 id=356 data-nosnippet>356</a> )
+<a href=#357 id=357 data-nosnippet>357</a> .<span class="kw">await
+<a href=#358 id=358 data-nosnippet>358</a> </span>.unwrap()
+<a href=#359 id=359 data-nosnippet>359</a> .unwrap();
<a href=#360 id=360 data-nosnippet>360</a>
-<a href=#361 id=361 data-nosnippet>361</a> </span><span class="kw">let
</span>delete_filter = delete_file_loader
-<a href=#362 id=362 data-nosnippet>362</a> .load_deletes(
-<a href=#363 id=363 data-nosnippet>363</a>
file_scan_tasks[<span class="number">1</span>].deletes(),
-<a href=#364 id=364 data-nosnippet>364</a>
file_scan_tasks[<span class="number">1</span>].schema_ref(),
-<a href=#365 id=365 data-nosnippet>365</a> )
-<a href=#366 id=366 data-nosnippet>366</a> .<span class="kw">await
-<a href=#367 id=367 data-nosnippet>367</a> </span>.unwrap()
-<a href=#368 id=368 data-nosnippet>368</a> .unwrap();
-<a href=#369 id=369 data-nosnippet>369</a>
-<a href=#370 id=370 data-nosnippet>370</a> <span class="kw">let
</span>result = delete_filter
-<a href=#371 id=371 data-nosnippet>371</a> .get_delete_vector(<span
class="kw-2">&</span>file_scan_tasks[<span class="number">1</span>])
-<a href=#372 id=372 data-nosnippet>372</a> .unwrap();
-<a href=#373 id=373 data-nosnippet>373</a> <span
class="macro">assert_eq!</span>(result.lock().unwrap().len(), <span
class="number">8</span>); <span class="comment">// no pos dels for file 3
-<a href=#374 id=374 data-nosnippet>374</a> </span>}
-<a href=#375 id=375 data-nosnippet>375</a>
-<a href=#376 id=376 data-nosnippet>376</a> <span
class="kw">pub</span>(<span class="kw">crate</span>) <span class="kw">fn
</span>setup(table_location: <span class="kw-2">&</span>Path) ->
Vec<FileScanTask> {
-<a href=#377 id=377 data-nosnippet>377</a> <span class="kw">let
</span>data_file_schema = Arc::new(Schema::builder().build().unwrap());
-<a href=#378 id=378 data-nosnippet>378</a> <span class="kw">let
</span>positional_delete_schema = create_pos_del_schema();
-<a href=#379 id=379 data-nosnippet>379</a>
-<a href=#380 id=380 data-nosnippet>380</a> <span class="kw">let
</span>file_path_values = [
-<a href=#381 id=381 data-nosnippet>381</a> <span
class="macro">vec!</span>[<span class="macro">format!</span>(<span
class="string">"{}/1.parquet"</span>, table_location.to_str().unwrap()); <span
class="number">8</span>],
-<a href=#382 id=382 data-nosnippet>382</a> <span
class="macro">vec!</span>[<span class="macro">format!</span>(<span
class="string">"{}/1.parquet"</span>, table_location.to_str().unwrap()); <span
class="number">8</span>],
-<a href=#383 id=383 data-nosnippet>383</a> <span
class="macro">vec!</span>[<span class="macro">format!</span>(<span
class="string">"{}/2.parquet"</span>, table_location.to_str().unwrap()); <span
class="number">8</span>],
-<a href=#384 id=384 data-nosnippet>384</a> ];
-<a href=#385 id=385 data-nosnippet>385</a> <span class="kw">let
</span>pos_values = [
-<a href=#386 id=386 data-nosnippet>386</a> <span
class="macro">vec!</span>[<span class="number">0i64</span>, <span
class="number">1</span>, <span class="number">3</span>, <span
class="number">5</span>, <span class="number">6</span>, <span
class="number">8</span>, <span class="number">1022</span>, <span
class="number">1023</span>],
-<a href=#387 id=387 data-nosnippet>387</a> <span
class="macro">vec!</span>[<span class="number">0i64</span>, <span
class="number">1</span>, <span class="number">3</span>, <span
class="number">5</span>, <span class="number">20</span>, <span
class="number">21</span>, <span class="number">22</span>, <span
class="number">23</span>],
-<a href=#388 id=388 data-nosnippet>388</a> <span
class="macro">vec!</span>[<span class="number">0i64</span>, <span
class="number">1</span>, <span class="number">3</span>, <span
class="number">5</span>, <span class="number">6</span>, <span
class="number">8</span>, <span class="number">1022</span>, <span
class="number">1023</span>],
+<a href=#361 id=361 data-nosnippet>361</a> <span class="kw">let
</span>result = delete_filter
+<a href=#362 id=362 data-nosnippet>362</a> .get_delete_vector(<span
class="kw-2">&</span>file_scan_tasks[<span class="number">0</span>])
+<a href=#363 id=363 data-nosnippet>363</a> .unwrap();
+<a href=#364 id=364 data-nosnippet>364</a> <span
class="macro">assert_eq!</span>(result.lock().unwrap().len(), <span
class="number">12</span>); <span class="comment">// pos dels from pos del file
1 and 2
+<a href=#365 id=365 data-nosnippet>365</a>
+<a href=#366 id=366 data-nosnippet>366</a> </span><span class="kw">let
</span>delete_filter = delete_file_loader
+<a href=#367 id=367 data-nosnippet>367</a> .load_deletes(
+<a href=#368 id=368 data-nosnippet>368</a>
file_scan_tasks[<span class="number">1</span>].deletes(),
+<a href=#369 id=369 data-nosnippet>369</a>
file_scan_tasks[<span class="number">1</span>].schema_ref(),
+<a href=#370 id=370 data-nosnippet>370</a> )
+<a href=#371 id=371 data-nosnippet>371</a> .<span class="kw">await
+<a href=#372 id=372 data-nosnippet>372</a> </span>.unwrap()
+<a href=#373 id=373 data-nosnippet>373</a> .unwrap();
+<a href=#374 id=374 data-nosnippet>374</a>
+<a href=#375 id=375 data-nosnippet>375</a> <span class="kw">let
</span>result = delete_filter
+<a href=#376 id=376 data-nosnippet>376</a> .get_delete_vector(<span
class="kw-2">&</span>file_scan_tasks[<span class="number">1</span>])
+<a href=#377 id=377 data-nosnippet>377</a> .unwrap();
+<a href=#378 id=378 data-nosnippet>378</a> <span
class="macro">assert_eq!</span>(result.lock().unwrap().len(), <span
class="number">8</span>); <span class="comment">// no pos dels for file 3
+<a href=#379 id=379 data-nosnippet>379</a> </span>}
+<a href=#380 id=380 data-nosnippet>380</a>
+<a href=#381 id=381 data-nosnippet>381</a> <span
class="kw">pub</span>(<span class="kw">crate</span>) <span class="kw">fn
</span>setup(table_location: <span class="kw-2">&</span>Path) ->
Vec<FileScanTask> {
+<a href=#382 id=382 data-nosnippet>382</a> <span class="kw">let
</span>data_file_schema = Arc::new(Schema::builder().build().unwrap());
+<a href=#383 id=383 data-nosnippet>383</a> <span class="kw">let
</span>positional_delete_schema = create_pos_del_schema();
+<a href=#384 id=384 data-nosnippet>384</a>
+<a href=#385 id=385 data-nosnippet>385</a> <span class="kw">let
</span>file_path_values = [
+<a href=#386 id=386 data-nosnippet>386</a> <span
class="macro">vec!</span>[<span class="macro">format!</span>(<span
class="string">"{}/1.parquet"</span>, table_location.to_str().unwrap()); <span
class="number">8</span>],
+<a href=#387 id=387 data-nosnippet>387</a> <span
class="macro">vec!</span>[<span class="macro">format!</span>(<span
class="string">"{}/1.parquet"</span>, table_location.to_str().unwrap()); <span
class="number">8</span>],
+<a href=#388 id=388 data-nosnippet>388</a> <span
class="macro">vec!</span>[<span class="macro">format!</span>(<span
class="string">"{}/2.parquet"</span>, table_location.to_str().unwrap()); <span
class="number">8</span>],
<a href=#389 id=389 data-nosnippet>389</a> ];
-<a href=#390 id=390 data-nosnippet>390</a>
-<a href=#391 id=391 data-nosnippet>391</a> <span class="kw">let
</span>props = WriterProperties::builder()
-<a href=#392 id=392 data-nosnippet>392</a>
.set_compression(Compression::SNAPPY)
-<a href=#393 id=393 data-nosnippet>393</a> .build();
-<a href=#394 id=394 data-nosnippet>394</a>
-<a href=#395 id=395 data-nosnippet>395</a> <span class="kw">for
</span>n <span class="kw">in </span><span class="number">1</span>..=<span
class="number">3 </span>{
-<a href=#396 id=396 data-nosnippet>396</a> <span class="kw">let
</span>file_path_vals = file_path_values.get(n - <span
class="number">1</span>).unwrap();
-<a href=#397 id=397 data-nosnippet>397</a> <span class="kw">let
</span>file_path_col = Arc::new(StringArray::from_iter_values(file_path_vals));
-<a href=#398 id=398 data-nosnippet>398</a>
-<a href=#399 id=399 data-nosnippet>399</a> <span class="kw">let
</span>pos_vals = pos_values.get(n - <span class="number">1</span>).unwrap();
-<a href=#400 id=400 data-nosnippet>400</a> <span class="kw">let
</span>pos_col = Arc::new(Int64Array::from_iter_values(pos_vals.clone()));
-<a href=#401 id=401 data-nosnippet>401</a>
-<a href=#402 id=402 data-nosnippet>402</a> <span class="kw">let
</span>positional_deletes_to_write =
-<a href=#403 id=403 data-nosnippet>403</a>
RecordBatch::try_new(positional_delete_schema.clone(), <span
class="macro">vec!</span>[
-<a href=#404 id=404 data-nosnippet>404</a>
file_path_col.clone(),
-<a href=#405 id=405 data-nosnippet>405</a> pos_col.clone(),
-<a href=#406 id=406 data-nosnippet>406</a> ])
-<a href=#407 id=407 data-nosnippet>407</a> .unwrap();
-<a href=#408 id=408 data-nosnippet>408</a>
-<a href=#409 id=409 data-nosnippet>409</a> <span class="kw">let
</span>file = File::create(<span class="macro">format!</span>(
-<a href=#410 id=410 data-nosnippet>410</a> <span
class="string">"{}/pos-del-{}.parquet"</span>,
-<a href=#411 id=411 data-nosnippet>411</a>
table_location.to_str().unwrap(),
-<a href=#412 id=412 data-nosnippet>412</a> n
-<a href=#413 id=413 data-nosnippet>413</a> ))
-<a href=#414 id=414 data-nosnippet>414</a> .unwrap();
-<a href=#415 id=415 data-nosnippet>415</a> <span class="kw">let
</span><span class="kw-2">mut </span>writer = ArrowWriter::try_new(
-<a href=#416 id=416 data-nosnippet>416</a> file,
-<a href=#417 id=417 data-nosnippet>417</a>
positional_deletes_to_write.schema(),
-<a href=#418 id=418 data-nosnippet>418</a> <span
class="prelude-val">Some</span>(props.clone()),
-<a href=#419 id=419 data-nosnippet>419</a> )
-<a href=#420 id=420 data-nosnippet>420</a> .unwrap();
-<a href=#421 id=421 data-nosnippet>421</a>
-<a href=#422 id=422 data-nosnippet>422</a> writer
-<a href=#423 id=423 data-nosnippet>423</a> .write(<span
class="kw-2">&</span>positional_deletes_to_write)
-<a href=#424 id=424 data-nosnippet>424</a> .expect(<span
class="string">"Writing batch"</span>);
-<a href=#425 id=425 data-nosnippet>425</a>
-<a href=#426 id=426 data-nosnippet>426</a> <span class="comment">//
writer must be closed to write footer
-<a href=#427 id=427 data-nosnippet>427</a>
</span>writer.close().unwrap();
-<a href=#428 id=428 data-nosnippet>428</a> }
-<a href=#429 id=429 data-nosnippet>429</a>
-<a href=#430 id=430 data-nosnippet>430</a> <span class="kw">let
</span>pos_del_1 = FileScanTaskDeleteFile::builder()
-<a href=#431 id=431 data-nosnippet>431</a> .with_file_path(<span
class="macro">format!</span>(
-<a href=#432 id=432 data-nosnippet>432</a> <span
class="string">"{}/pos-del-1.parquet"</span>,
-<a href=#433 id=433 data-nosnippet>433</a>
table_location.to_str().unwrap()
-<a href=#434 id=434 data-nosnippet>434</a> ))
-<a href=#435 id=435 data-nosnippet>435</a> .with_file_size_in_bytes(
-<a href=#436 id=436 data-nosnippet>436</a>
std::fs::metadata(<span class="macro">format!</span>(
-<a href=#437 id=437 data-nosnippet>437</a> <span
class="string">"{}/pos-del-1.parquet"</span>,
-<a href=#438 id=438 data-nosnippet>438</a>
table_location.to_str().unwrap()
-<a href=#439 id=439 data-nosnippet>439</a> ))
-<a href=#440 id=440 data-nosnippet>440</a> .unwrap()
-<a href=#441 id=441 data-nosnippet>441</a> .len(),
-<a href=#442 id=442 data-nosnippet>442</a> )
-<a href=#443 id=443 data-nosnippet>443</a>
.with_file_type(DataContentType::PositionDeletes)
-<a href=#444 id=444 data-nosnippet>444</a>
.with_file_format(DataFileFormat::Parquet)
-<a href=#445 id=445 data-nosnippet>445</a>
.with_partition_spec_id(<span class="number">0</span>)
-<a href=#446 id=446 data-nosnippet>446</a> .build();
-<a href=#447 id=447 data-nosnippet>447</a>
-<a href=#448 id=448 data-nosnippet>448</a> <span class="kw">let
</span>pos_del_2 = FileScanTaskDeleteFile::builder()
-<a href=#449 id=449 data-nosnippet>449</a> .with_file_path(<span
class="macro">format!</span>(
-<a href=#450 id=450 data-nosnippet>450</a> <span
class="string">"{}/pos-del-2.parquet"</span>,
-<a href=#451 id=451 data-nosnippet>451</a>
table_location.to_str().unwrap()
-<a href=#452 id=452 data-nosnippet>452</a> ))
-<a href=#453 id=453 data-nosnippet>453</a> .with_file_size_in_bytes(
-<a href=#454 id=454 data-nosnippet>454</a>
std::fs::metadata(<span class="macro">format!</span>(
-<a href=#455 id=455 data-nosnippet>455</a> <span
class="string">"{}/pos-del-2.parquet"</span>,
-<a href=#456 id=456 data-nosnippet>456</a>
table_location.to_str().unwrap()
-<a href=#457 id=457 data-nosnippet>457</a> ))
-<a href=#458 id=458 data-nosnippet>458</a> .unwrap()
-<a href=#459 id=459 data-nosnippet>459</a> .len(),
-<a href=#460 id=460 data-nosnippet>460</a> )
-<a href=#461 id=461 data-nosnippet>461</a>
.with_file_type(DataContentType::PositionDeletes)
-<a href=#462 id=462 data-nosnippet>462</a>
.with_file_format(DataFileFormat::Parquet)
-<a href=#463 id=463 data-nosnippet>463</a>
.with_partition_spec_id(<span class="number">0</span>)
-<a href=#464 id=464 data-nosnippet>464</a> .build();
-<a href=#465 id=465 data-nosnippet>465</a>
-<a href=#466 id=466 data-nosnippet>466</a> <span class="kw">let
</span>pos_del_3 = FileScanTaskDeleteFile::builder()
-<a href=#467 id=467 data-nosnippet>467</a> .with_file_path(<span
class="macro">format!</span>(
-<a href=#468 id=468 data-nosnippet>468</a> <span
class="string">"{}/pos-del-3.parquet"</span>,
-<a href=#469 id=469 data-nosnippet>469</a>
table_location.to_str().unwrap()
-<a href=#470 id=470 data-nosnippet>470</a> ))
-<a href=#471 id=471 data-nosnippet>471</a> .with_file_size_in_bytes(
-<a href=#472 id=472 data-nosnippet>472</a>
std::fs::metadata(<span class="macro">format!</span>(
-<a href=#473 id=473 data-nosnippet>473</a> <span
class="string">"{}/pos-del-3.parquet"</span>,
-<a href=#474 id=474 data-nosnippet>474</a>
table_location.to_str().unwrap()
-<a href=#475 id=475 data-nosnippet>475</a> ))
-<a href=#476 id=476 data-nosnippet>476</a> .unwrap()
-<a href=#477 id=477 data-nosnippet>477</a> .len(),
-<a href=#478 id=478 data-nosnippet>478</a> )
-<a href=#479 id=479 data-nosnippet>479</a>
.with_file_type(DataContentType::PositionDeletes)
-<a href=#480 id=480 data-nosnippet>480</a>
.with_file_format(DataFileFormat::Parquet)
-<a href=#481 id=481 data-nosnippet>481</a>
.with_partition_spec_id(<span class="number">0</span>)
-<a href=#482 id=482 data-nosnippet>482</a> .build();
-<a href=#483 id=483 data-nosnippet>483</a>
-<a href=#484 id=484 data-nosnippet>484</a> <span class="kw">let
</span>file_scan_tasks = <span class="macro">vec!</span>[
-<a href=#485 id=485 data-nosnippet>485</a> FileScanTask::builder()
-<a href=#486 id=486 data-nosnippet>486</a>
.with_file_size_in_bytes(<span class="number">0</span>)
-<a href=#487 id=487 data-nosnippet>487</a> .with_start(<span
class="number">0</span>)
-<a href=#488 id=488 data-nosnippet>488</a> .with_length(<span
class="number">0</span>)
-<a href=#489 id=489 data-nosnippet>489</a>
.with_data_file_path(<span class="macro">format!</span>(<span
class="string">"{}/1.parquet"</span>, table_location.to_str().unwrap()))
-<a href=#490 id=490 data-nosnippet>490</a>
.with_data_file_format(DataFileFormat::Parquet)
-<a href=#491 id=491 data-nosnippet>491</a>
.with_schema(data_file_schema.clone())
-<a href=#492 id=492 data-nosnippet>492</a>
.with_project_field_ids(<span class="macro">vec!</span>[])
-<a href=#493 id=493 data-nosnippet>493</a> .with_deletes(<span
class="macro">vec!</span>[pos_del_1, pos_del_2.clone()])
-<a href=#494 id=494 data-nosnippet>494</a>
.with_case_sensitive(<span class="bool-val">false</span>)
-<a href=#495 id=495 data-nosnippet>495</a> .build()
-<a href=#496 id=496 data-nosnippet>496</a> .unwrap(),
-<a href=#497 id=497 data-nosnippet>497</a> FileScanTask::builder()
-<a href=#498 id=498 data-nosnippet>498</a>
.with_file_size_in_bytes(<span class="number">0</span>)
-<a href=#499 id=499 data-nosnippet>499</a> .with_start(<span
class="number">0</span>)
-<a href=#500 id=500 data-nosnippet>500</a> .with_length(<span
class="number">0</span>)
-<a href=#501 id=501 data-nosnippet>501</a>
.with_data_file_path(<span class="macro">format!</span>(<span
class="string">"{}/2.parquet"</span>, table_location.to_str().unwrap()))
-<a href=#502 id=502 data-nosnippet>502</a>
.with_data_file_format(DataFileFormat::Parquet)
-<a href=#503 id=503 data-nosnippet>503</a>
.with_schema(data_file_schema.clone())
-<a href=#504 id=504 data-nosnippet>504</a>
.with_project_field_ids(<span class="macro">vec!</span>[])
-<a href=#505 id=505 data-nosnippet>505</a> .with_deletes(<span
class="macro">vec!</span>[pos_del_3])
-<a href=#506 id=506 data-nosnippet>506</a>
.with_case_sensitive(<span class="bool-val">false</span>)
-<a href=#507 id=507 data-nosnippet>507</a> .build()
-<a href=#508 id=508 data-nosnippet>508</a> .unwrap(),
-<a href=#509 id=509 data-nosnippet>509</a> ];
-<a href=#510 id=510 data-nosnippet>510</a>
-<a href=#511 id=511 data-nosnippet>511</a> file_scan_tasks
-<a href=#512 id=512 data-nosnippet>512</a> }
-<a href=#513 id=513 data-nosnippet>513</a>
-<a href=#514 id=514 data-nosnippet>514</a> <span
class="kw">pub</span>(<span class="kw">crate</span>) <span class="kw">fn
</span>create_pos_del_schema() -> ArrowSchemaRef {
-<a href=#515 id=515 data-nosnippet>515</a> <span class="kw">let
</span>fields = <span class="macro">vec!</span>[
-<a href=#516 id=516 data-nosnippet>516</a>
arrow_schema::Field::new(<span class="string">"file_path"</span>,
arrow_schema::DataType::Utf8, <span class="bool-val">false</span>)
-<a href=#517 id=517 data-nosnippet>517</a>
.with_metadata(HashMap::from([(
-<a href=#518 id=518 data-nosnippet>518</a>
PARQUET_FIELD_ID_META_KEY.to_string(),
-<a href=#519 id=519 data-nosnippet>519</a>
FIELD_ID_POSITIONAL_DELETE_FILE_PATH.to_string(),
-<a href=#520 id=520 data-nosnippet>520</a> )])),
-<a href=#521 id=521 data-nosnippet>521</a>
arrow_schema::Field::new(<span class="string">"pos"</span>,
arrow_schema::DataType::Int64, <span
class="bool-val">false</span>).with_metadata(
-<a href=#522 id=522 data-nosnippet>522</a> HashMap::from([(
+<a href=#390 id=390 data-nosnippet>390</a> <span class="kw">let
</span>pos_values = [
+<a href=#391 id=391 data-nosnippet>391</a> <span
class="macro">vec!</span>[<span class="number">0i64</span>, <span
class="number">1</span>, <span class="number">3</span>, <span
class="number">5</span>, <span class="number">6</span>, <span
class="number">8</span>, <span class="number">1022</span>, <span
class="number">1023</span>],
+<a href=#392 id=392 data-nosnippet>392</a> <span
class="macro">vec!</span>[<span class="number">0i64</span>, <span
class="number">1</span>, <span class="number">3</span>, <span
class="number">5</span>, <span class="number">20</span>, <span
class="number">21</span>, <span class="number">22</span>, <span
class="number">23</span>],
+<a href=#393 id=393 data-nosnippet>393</a> <span
class="macro">vec!</span>[<span class="number">0i64</span>, <span
class="number">1</span>, <span class="number">3</span>, <span
class="number">5</span>, <span class="number">6</span>, <span
class="number">8</span>, <span class="number">1022</span>, <span
class="number">1023</span>],
+<a href=#394 id=394 data-nosnippet>394</a> ];
+<a href=#395 id=395 data-nosnippet>395</a>
+<a href=#396 id=396 data-nosnippet>396</a> <span class="kw">let
</span>props = WriterProperties::builder()
+<a href=#397 id=397 data-nosnippet>397</a>
.set_compression(Compression::SNAPPY)
+<a href=#398 id=398 data-nosnippet>398</a> .build();
+<a href=#399 id=399 data-nosnippet>399</a>
+<a href=#400 id=400 data-nosnippet>400</a> <span class="kw">for
</span>n <span class="kw">in </span><span class="number">1</span>..=<span
class="number">3 </span>{
+<a href=#401 id=401 data-nosnippet>401</a> <span class="kw">let
</span>file_path_vals = file_path_values.get(n - <span
class="number">1</span>).unwrap();
+<a href=#402 id=402 data-nosnippet>402</a> <span class="kw">let
</span>file_path_col = Arc::new(StringArray::from_iter_values(file_path_vals));
+<a href=#403 id=403 data-nosnippet>403</a>
+<a href=#404 id=404 data-nosnippet>404</a> <span class="kw">let
</span>pos_vals = pos_values.get(n - <span class="number">1</span>).unwrap();
+<a href=#405 id=405 data-nosnippet>405</a> <span class="kw">let
</span>pos_col = Arc::new(Int64Array::from_iter_values(pos_vals.clone()));
+<a href=#406 id=406 data-nosnippet>406</a>
+<a href=#407 id=407 data-nosnippet>407</a> <span class="kw">let
</span>positional_deletes_to_write =
+<a href=#408 id=408 data-nosnippet>408</a>
RecordBatch::try_new(positional_delete_schema.clone(), <span
class="macro">vec!</span>[
+<a href=#409 id=409 data-nosnippet>409</a>
file_path_col.clone(),
+<a href=#410 id=410 data-nosnippet>410</a> pos_col.clone(),
+<a href=#411 id=411 data-nosnippet>411</a> ])
+<a href=#412 id=412 data-nosnippet>412</a> .unwrap();
+<a href=#413 id=413 data-nosnippet>413</a>
+<a href=#414 id=414 data-nosnippet>414</a> <span class="kw">let
</span>file = File::create(<span class="macro">format!</span>(
+<a href=#415 id=415 data-nosnippet>415</a> <span
class="string">"{}/pos-del-{}.parquet"</span>,
+<a href=#416 id=416 data-nosnippet>416</a>
table_location.to_str().unwrap(),
+<a href=#417 id=417 data-nosnippet>417</a> n
+<a href=#418 id=418 data-nosnippet>418</a> ))
+<a href=#419 id=419 data-nosnippet>419</a> .unwrap();
+<a href=#420 id=420 data-nosnippet>420</a> <span class="kw">let
</span><span class="kw-2">mut </span>writer = ArrowWriter::try_new(
+<a href=#421 id=421 data-nosnippet>421</a> file,
+<a href=#422 id=422 data-nosnippet>422</a>
positional_deletes_to_write.schema(),
+<a href=#423 id=423 data-nosnippet>423</a> <span
class="prelude-val">Some</span>(props.clone()),
+<a href=#424 id=424 data-nosnippet>424</a> )
+<a href=#425 id=425 data-nosnippet>425</a> .unwrap();
+<a href=#426 id=426 data-nosnippet>426</a>
+<a href=#427 id=427 data-nosnippet>427</a> writer
+<a href=#428 id=428 data-nosnippet>428</a> .write(<span
class="kw-2">&</span>positional_deletes_to_write)
+<a href=#429 id=429 data-nosnippet>429</a> .expect(<span
class="string">"Writing batch"</span>);
+<a href=#430 id=430 data-nosnippet>430</a>
+<a href=#431 id=431 data-nosnippet>431</a> <span class="comment">//
writer must be closed to write footer
+<a href=#432 id=432 data-nosnippet>432</a>
</span>writer.close().unwrap();
+<a href=#433 id=433 data-nosnippet>433</a> }
+<a href=#434 id=434 data-nosnippet>434</a>
+<a href=#435 id=435 data-nosnippet>435</a> <span class="kw">let
</span>pos_del_1 = FileScanTaskDeleteFile::builder()
+<a href=#436 id=436 data-nosnippet>436</a> .with_file_path(<span
class="macro">format!</span>(
+<a href=#437 id=437 data-nosnippet>437</a> <span
class="string">"{}/pos-del-1.parquet"</span>,
+<a href=#438 id=438 data-nosnippet>438</a>
table_location.to_str().unwrap()
+<a href=#439 id=439 data-nosnippet>439</a> ))
+<a href=#440 id=440 data-nosnippet>440</a> .with_file_size_in_bytes(
+<a href=#441 id=441 data-nosnippet>441</a>
std::fs::metadata(<span class="macro">format!</span>(
+<a href=#442 id=442 data-nosnippet>442</a> <span
class="string">"{}/pos-del-1.parquet"</span>,
+<a href=#443 id=443 data-nosnippet>443</a>
table_location.to_str().unwrap()
+<a href=#444 id=444 data-nosnippet>444</a> ))
+<a href=#445 id=445 data-nosnippet>445</a> .unwrap()
+<a href=#446 id=446 data-nosnippet>446</a> .len(),
+<a href=#447 id=447 data-nosnippet>447</a> )
+<a href=#448 id=448 data-nosnippet>448</a>
.with_file_type(DataContentType::PositionDeletes)
+<a href=#449 id=449 data-nosnippet>449</a>
.with_file_format(DataFileFormat::Parquet)
+<a href=#450 id=450 data-nosnippet>450</a>
.with_partition_spec_id(<span class="number">0</span>)
+<a href=#451 id=451 data-nosnippet>451</a> .build();
+<a href=#452 id=452 data-nosnippet>452</a>
+<a href=#453 id=453 data-nosnippet>453</a> <span class="kw">let
</span>pos_del_2 = FileScanTaskDeleteFile::builder()
+<a href=#454 id=454 data-nosnippet>454</a> .with_file_path(<span
class="macro">format!</span>(
+<a href=#455 id=455 data-nosnippet>455</a> <span
class="string">"{}/pos-del-2.parquet"</span>,
+<a href=#456 id=456 data-nosnippet>456</a>
table_location.to_str().unwrap()
+<a href=#457 id=457 data-nosnippet>457</a> ))
+<a href=#458 id=458 data-nosnippet>458</a> .with_file_size_in_bytes(
+<a href=#459 id=459 data-nosnippet>459</a>
std::fs::metadata(<span class="macro">format!</span>(
+<a href=#460 id=460 data-nosnippet>460</a> <span
class="string">"{}/pos-del-2.parquet"</span>,
+<a href=#461 id=461 data-nosnippet>461</a>
table_location.to_str().unwrap()
+<a href=#462 id=462 data-nosnippet>462</a> ))
+<a href=#463 id=463 data-nosnippet>463</a> .unwrap()
+<a href=#464 id=464 data-nosnippet>464</a> .len(),
+<a href=#465 id=465 data-nosnippet>465</a> )
+<a href=#466 id=466 data-nosnippet>466</a>
.with_file_type(DataContentType::PositionDeletes)
+<a href=#467 id=467 data-nosnippet>467</a>
.with_file_format(DataFileFormat::Parquet)
+<a href=#468 id=468 data-nosnippet>468</a>
.with_partition_spec_id(<span class="number">0</span>)
+<a href=#469 id=469 data-nosnippet>469</a> .build();
+<a href=#470 id=470 data-nosnippet>470</a>
+<a href=#471 id=471 data-nosnippet>471</a> <span class="kw">let
</span>pos_del_3 = FileScanTaskDeleteFile::builder()
+<a href=#472 id=472 data-nosnippet>472</a> .with_file_path(<span
class="macro">format!</span>(
+<a href=#473 id=473 data-nosnippet>473</a> <span
class="string">"{}/pos-del-3.parquet"</span>,
+<a href=#474 id=474 data-nosnippet>474</a>
table_location.to_str().unwrap()
+<a href=#475 id=475 data-nosnippet>475</a> ))
+<a href=#476 id=476 data-nosnippet>476</a> .with_file_size_in_bytes(
+<a href=#477 id=477 data-nosnippet>477</a>
std::fs::metadata(<span class="macro">format!</span>(
+<a href=#478 id=478 data-nosnippet>478</a> <span
class="string">"{}/pos-del-3.parquet"</span>,
+<a href=#479 id=479 data-nosnippet>479</a>
table_location.to_str().unwrap()
+<a href=#480 id=480 data-nosnippet>480</a> ))
+<a href=#481 id=481 data-nosnippet>481</a> .unwrap()
+<a href=#482 id=482 data-nosnippet>482</a> .len(),
+<a href=#483 id=483 data-nosnippet>483</a> )
+<a href=#484 id=484 data-nosnippet>484</a>
.with_file_type(DataContentType::PositionDeletes)
+<a href=#485 id=485 data-nosnippet>485</a>
.with_file_format(DataFileFormat::Parquet)
+<a href=#486 id=486 data-nosnippet>486</a>
.with_partition_spec_id(<span class="number">0</span>)
+<a href=#487 id=487 data-nosnippet>487</a> .build();
+<a href=#488 id=488 data-nosnippet>488</a>
+<a href=#489 id=489 data-nosnippet>489</a> <span class="kw">let
</span>file_scan_tasks = <span class="macro">vec!</span>[
+<a href=#490 id=490 data-nosnippet>490</a> FileScanTask::builder()
+<a href=#491 id=491 data-nosnippet>491</a>
.with_file_size_in_bytes(<span class="number">0</span>)
+<a href=#492 id=492 data-nosnippet>492</a> .with_start(<span
class="number">0</span>)
+<a href=#493 id=493 data-nosnippet>493</a> .with_length(<span
class="number">0</span>)
+<a href=#494 id=494 data-nosnippet>494</a>
.with_data_file_path(<span class="macro">format!</span>(<span
class="string">"{}/1.parquet"</span>, table_location.to_str().unwrap()))
+<a href=#495 id=495 data-nosnippet>495</a>
.with_data_file_format(DataFileFormat::Parquet)
+<a href=#496 id=496 data-nosnippet>496</a>
.with_schema(data_file_schema.clone())
+<a href=#497 id=497 data-nosnippet>497</a>
.with_project_field_ids(<span class="macro">vec!</span>[])
+<a href=#498 id=498 data-nosnippet>498</a> .with_deletes(<span
class="macro">vec!</span>[pos_del_1, pos_del_2.clone()])
+<a href=#499 id=499 data-nosnippet>499</a>
.with_case_sensitive(<span class="bool-val">false</span>)
+<a href=#500 id=500 data-nosnippet>500</a> .build()
+<a href=#501 id=501 data-nosnippet>501</a> .unwrap(),
+<a href=#502 id=502 data-nosnippet>502</a> FileScanTask::builder()
+<a href=#503 id=503 data-nosnippet>503</a>
.with_file_size_in_bytes(<span class="number">0</span>)
+<a href=#504 id=504 data-nosnippet>504</a> .with_start(<span
class="number">0</span>)
+<a href=#505 id=505 data-nosnippet>505</a> .with_length(<span
class="number">0</span>)
+<a href=#506 id=506 data-nosnippet>506</a>
.with_data_file_path(<span class="macro">format!</span>(<span
class="string">"{}/2.parquet"</span>, table_location.to_str().unwrap()))
+<a href=#507 id=507 data-nosnippet>507</a>
.with_data_file_format(DataFileFormat::Parquet)
+<a href=#508 id=508 data-nosnippet>508</a>
.with_schema(data_file_schema.clone())
+<a href=#509 id=509 data-nosnippet>509</a>
.with_project_field_ids(<span class="macro">vec!</span>[])
+<a href=#510 id=510 data-nosnippet>510</a> .with_deletes(<span
class="macro">vec!</span>[pos_del_3])
+<a href=#511 id=511 data-nosnippet>511</a>
.with_case_sensitive(<span class="bool-val">false</span>)
+<a href=#512 id=512 data-nosnippet>512</a> .build()
+<a href=#513 id=513 data-nosnippet>513</a> .unwrap(),
+<a href=#514 id=514 data-nosnippet>514</a> ];
+<a href=#515 id=515 data-nosnippet>515</a>
+<a href=#516 id=516 data-nosnippet>516</a> file_scan_tasks
+<a href=#517 id=517 data-nosnippet>517</a> }
+<a href=#518 id=518 data-nosnippet>518</a>
+<a href=#519 id=519 data-nosnippet>519</a> <span
class="kw">pub</span>(<span class="kw">crate</span>) <span class="kw">fn
</span>create_pos_del_schema() -> ArrowSchemaRef {
+<a href=#520 id=520 data-nosnippet>520</a> <span class="kw">let
</span>fields = <span class="macro">vec!</span>[
+<a href=#521 id=521 data-nosnippet>521</a>
arrow_schema::Field::new(<span class="string">"file_path"</span>,
arrow_schema::DataType::Utf8, <span class="bool-val">false</span>)
+<a href=#522 id=522 data-nosnippet>522</a>
.with_metadata(HashMap::from([(
<a href=#523 id=523 data-nosnippet>523</a>
PARQUET_FIELD_ID_META_KEY.to_string(),
-<a href=#524 id=524 data-nosnippet>524</a>
FIELD_ID_POSITIONAL_DELETE_POS.to_string(),
-<a href=#525 id=525 data-nosnippet>525</a> )]),
-<a href=#526 id=526 data-nosnippet>526</a> ),
-<a href=#527 id=527 data-nosnippet>527</a> ];
-<a href=#528 id=528 data-nosnippet>528</a>
Arc::new(arrow_schema::Schema::new(fields))
-<a href=#529 id=529 data-nosnippet>529</a> }
-<a href=#530 id=530 data-nosnippet>530</a>
-<a href=#531 id=531 data-nosnippet>531</a> <span class="attr">#[tokio::test]
-<a href=#532 id=532 data-nosnippet>532</a> </span><span class="kw">async fn
</span>test_build_equality_delete_predicate_case_sensitive() {
-<a href=#533 id=533 data-nosnippet>533</a> <span class="kw">let
</span>schema = Arc::new(
-<a href=#534 id=534 data-nosnippet>534</a> Schema::builder()
-<a href=#535 id=535 data-nosnippet>535</a>
.with_schema_id(<span class="number">1</span>)
-<a href=#536 id=536 data-nosnippet>536</a> .with_fields(<span
class="macro">vec!</span>[
-<a href=#537 id=537 data-nosnippet>537</a>
NestedField::required(<span class="number">1</span>, <span
class="string">"Id"</span>, Type::Primitive(PrimitiveType::Long)).into(),
-<a href=#538 id=538 data-nosnippet>538</a> ])
-<a href=#539 id=539 data-nosnippet>539</a> .build()
-<a href=#540 id=540 data-nosnippet>540</a> .unwrap(),
-<a href=#541 id=541 data-nosnippet>541</a> );
-<a href=#542 id=542 data-nosnippet>542</a>
-<a href=#543 id=543 data-nosnippet>543</a> <span class="comment">//
---------- fake FileScanTask ----------
-<a href=#544 id=544 data-nosnippet>544</a> </span><span class="kw">let
</span>task = FileScanTask::builder()
-<a href=#545 id=545 data-nosnippet>545</a>
.with_file_size_in_bytes(<span class="number">0</span>)
-<a href=#546 id=546 data-nosnippet>546</a> .with_start(<span
class="number">0</span>)
-<a href=#547 id=547 data-nosnippet>547</a> .with_length(<span
class="number">0</span>)
-<a href=#548 id=548 data-nosnippet>548</a>
.with_data_file_path(<span class="string">"data.parquet"</span>.to_string())
-<a href=#549 id=549 data-nosnippet>549</a>
.with_data_file_format(DataFileFormat::Parquet)
-<a href=#550 id=550 data-nosnippet>550</a>
.with_schema(schema.clone())
-<a href=#551 id=551 data-nosnippet>551</a>
.with_project_field_ids(<span class="macro">vec!</span>[])
-<a href=#552 id=552 data-nosnippet>552</a> .with_deletes(<span
class="macro">vec!</span>[
-<a href=#553 id=553 data-nosnippet>553</a>
FileScanTaskDeleteFile::builder()
-<a href=#554 id=554 data-nosnippet>554</a>
.with_file_path(<span class="string">"eq-del.parquet"</span>.to_string())
-<a href=#555 id=555 data-nosnippet>555</a>
.with_file_size_in_bytes(<span class="number">1</span>) <span
class="comment">// never read; this test fails before opening the file
-<a href=#556 id=556 data-nosnippet>556</a>
</span>.with_file_type(DataContentType::EqualityDeletes)
-<a href=#557 id=557 data-nosnippet>557</a>
.with_file_format(DataFileFormat::Parquet)
-<a href=#558 id=558 data-nosnippet>558</a>
.with_partition_spec_id(<span class="number">0</span>)
-<a href=#559 id=559 data-nosnippet>559</a> .build(),
-<a href=#560 id=560 data-nosnippet>560</a> ])
-<a href=#561 id=561 data-nosnippet>561</a>
.with_case_sensitive(<span class="bool-val">true</span>)
-<a href=#562 id=562 data-nosnippet>562</a> .build()
-<a href=#563 id=563 data-nosnippet>563</a> .unwrap();
-<a href=#564 id=564 data-nosnippet>564</a>
-<a href=#565 id=565 data-nosnippet>565</a> <span class="kw">let
</span>filter = DeleteFilter::new(Runtime::current());
-<a href=#566 id=566 data-nosnippet>566</a>
-<a href=#567 id=567 data-nosnippet>567</a> <span class="comment">//
---------- insert equality delete predicate ----------
-<a href=#568 id=568 data-nosnippet>568</a> </span><span class="kw">let
</span>pred = Reference::new(<span
class="string">"id"</span>).equal_to(Datum::long(<span
class="number">10</span>));
+<a href=#524 id=524 data-nosnippet>524</a>
FIELD_ID_POSITIONAL_DELETE_FILE_PATH.to_string(),
+<a href=#525 id=525 data-nosnippet>525</a> )])),
+<a href=#526 id=526 data-nosnippet>526</a>
arrow_schema::Field::new(<span class="string">"pos"</span>,
arrow_schema::DataType::Int64, <span
class="bool-val">false</span>).with_metadata(
+<a href=#527 id=527 data-nosnippet>527</a> HashMap::from([(
+<a href=#528 id=528 data-nosnippet>528</a>
PARQUET_FIELD_ID_META_KEY.to_string(),
+<a href=#529 id=529 data-nosnippet>529</a>
FIELD_ID_POSITIONAL_DELETE_POS.to_string(),
+<a href=#530 id=530 data-nosnippet>530</a> )]),
+<a href=#531 id=531 data-nosnippet>531</a> ),
+<a href=#532 id=532 data-nosnippet>532</a> ];
+<a href=#533 id=533 data-nosnippet>533</a>
Arc::new(arrow_schema::Schema::new(fields))
+<a href=#534 id=534 data-nosnippet>534</a> }
+<a href=#535 id=535 data-nosnippet>535</a>
+<a href=#536 id=536 data-nosnippet>536</a> <span class="attr">#[tokio::test]
+<a href=#537 id=537 data-nosnippet>537</a> </span><span class="kw">async fn
</span>test_build_equality_delete_predicate_case_sensitive() {
+<a href=#538 id=538 data-nosnippet>538</a> <span class="kw">let
</span>schema = Arc::new(
+<a href=#539 id=539 data-nosnippet>539</a> Schema::builder()
+<a href=#540 id=540 data-nosnippet>540</a>
.with_schema_id(<span class="number">1</span>)
+<a href=#541 id=541 data-nosnippet>541</a> .with_fields(<span
class="macro">vec!</span>[
+<a href=#542 id=542 data-nosnippet>542</a>
NestedField::required(<span class="number">1</span>, <span
class="string">"Id"</span>, Type::Primitive(PrimitiveType::Long)).into(),
+<a href=#543 id=543 data-nosnippet>543</a> ])
+<a href=#544 id=544 data-nosnippet>544</a> .build()
+<a href=#545 id=545 data-nosnippet>545</a> .unwrap(),
+<a href=#546 id=546 data-nosnippet>546</a> );
+<a href=#547 id=547 data-nosnippet>547</a>
+<a href=#548 id=548 data-nosnippet>548</a> <span class="comment">//
---------- fake FileScanTask ----------
+<a href=#549 id=549 data-nosnippet>549</a> </span><span class="kw">let
</span>task = FileScanTask::builder()
+<a href=#550 id=550 data-nosnippet>550</a>
.with_file_size_in_bytes(<span class="number">0</span>)
+<a href=#551 id=551 data-nosnippet>551</a> .with_start(<span
class="number">0</span>)
+<a href=#552 id=552 data-nosnippet>552</a> .with_length(<span
class="number">0</span>)
+<a href=#553 id=553 data-nosnippet>553</a>
.with_data_file_path(<span class="string">"data.parquet"</span>.to_string())
+<a href=#554 id=554 data-nosnippet>554</a>
.with_data_file_format(DataFileFormat::Parquet)
+<a href=#555 id=555 data-nosnippet>555</a>
.with_schema(schema.clone())
+<a href=#556 id=556 data-nosnippet>556</a>
.with_project_field_ids(<span class="macro">vec!</span>[])
+<a href=#557 id=557 data-nosnippet>557</a> .with_deletes(<span
class="macro">vec!</span>[
+<a href=#558 id=558 data-nosnippet>558</a>
FileScanTaskDeleteFile::builder()
+<a href=#559 id=559 data-nosnippet>559</a>
.with_file_path(<span class="string">"eq-del.parquet"</span>.to_string())
+<a href=#560 id=560 data-nosnippet>560</a>
.with_file_size_in_bytes(<span class="number">1</span>) <span
class="comment">// never read; this test fails before opening the file
+<a href=#561 id=561 data-nosnippet>561</a>
</span>.with_file_type(DataContentType::EqualityDeletes)
+<a href=#562 id=562 data-nosnippet>562</a>
.with_file_format(DataFileFormat::Parquet)
+<a href=#563 id=563 data-nosnippet>563</a>
.with_partition_spec_id(<span class="number">0</span>)
+<a href=#564 id=564 data-nosnippet>564</a> .build(),
+<a href=#565 id=565 data-nosnippet>565</a> ])
+<a href=#566 id=566 data-nosnippet>566</a>
.with_case_sensitive(<span class="bool-val">true</span>)
+<a href=#567 id=567 data-nosnippet>567</a> .build()
+<a href=#568 id=568 data-nosnippet>568</a> .unwrap();
<a href=#569 id=569 data-nosnippet>569</a>
-<a href=#570 id=570 data-nosnippet>570</a> <span class="kw">let
</span>(tx, rx) = tokio::sync::oneshot::channel();
-<a href=#571 id=571 data-nosnippet>571</a>
filter.insert_equality_delete(<span class="string">"eq-del.parquet"</span>, rx);
-<a href=#572 id=572 data-nosnippet>572</a>
-<a href=#573 id=573 data-nosnippet>573</a> tx.send(pred).unwrap();
+<a href=#570 id=570 data-nosnippet>570</a> <span class="kw">let
</span>filter = DeleteFilter::new(Runtime::current());
+<a href=#571 id=571 data-nosnippet>571</a>
+<a href=#572 id=572 data-nosnippet>572</a> <span class="comment">//
---------- insert equality delete predicate ----------
+<a href=#573 id=573 data-nosnippet>573</a> </span><span class="kw">let
</span>pred = Reference::new(<span
class="string">"id"</span>).equal_to(Datum::long(<span
class="number">10</span>));
<a href=#574 id=574 data-nosnippet>574</a>
-<a href=#575 id=575 data-nosnippet>575</a> <span class="comment">//
---------- should FAIL ----------
-<a href=#576 id=576 data-nosnippet>576</a> </span><span class="kw">let
</span>result = filter.build_equality_delete_predicate(<span
class="kw-2">&</span>task).<span class="kw">await</span>;
+<a href=#575 id=575 data-nosnippet>575</a> <span class="kw">let
</span>(tx, rx) = tokio::sync::oneshot::channel();
+<a href=#576 id=576 data-nosnippet>576</a>
filter.insert_equality_delete(<span class="string">"eq-del.parquet"</span>, rx);
<a href=#577 id=577 data-nosnippet>577</a>
-<a href=#578 id=578 data-nosnippet>578</a> <span
class="macro">assert!</span>(
-<a href=#579 id=579 data-nosnippet>579</a> result.is_err(),
-<a href=#580 id=580 data-nosnippet>580</a> <span
class="string">"case_sensitive=true should fail when column case mismatches"
-<a href=#581 id=581 data-nosnippet>581</a> </span>);
-<a href=#582 id=582 data-nosnippet>582</a> }
-<a href=#583 id=583 data-nosnippet>583</a>}
+<a href=#578 id=578 data-nosnippet>578</a> tx.send(pred).unwrap();
+<a href=#579 id=579 data-nosnippet>579</a>
+<a href=#580 id=580 data-nosnippet>580</a> <span class="comment">//
---------- should FAIL ----------
+<a href=#581 id=581 data-nosnippet>581</a> </span><span class="kw">let
</span>result = filter.build_equality_delete_predicate(<span
class="kw-2">&</span>task).<span class="kw">await</span>;
+<a href=#582 id=582 data-nosnippet>582</a>
+<a href=#583 id=583 data-nosnippet>583</a> <span
class="macro">assert!</span>(
+<a href=#584 id=584 data-nosnippet>584</a> result.is_err(),
+<a href=#585 id=585 data-nosnippet>585</a> <span
class="string">"case_sensitive=true should fail when column case mismatches"
+<a href=#586 id=586 data-nosnippet>586</a> </span>);
+<a href=#587 id=587 data-nosnippet>587</a> }
+<a href=#588 id=588 data-nosnippet>588</a>}
</code></pre></div></section></main></body></html>
\ No newline at end of file
