#general


@working1634: @working1634 has joined the channel
@diana.arnos: Hello everyone :smile: Out of curiosity, do you have any idea when the next version will be released? :eyes:
  @atri.sharma: Its in process as we speak
  @diana.arnos: Thanks :heart:
@wcxzjtz: hey I was trying to use geoindex feature in pinot. but seems the index doesn’t apply because the `numEntriesScannedInFilter` . is high (equals to the number of docs). the pinot version is `0.8.0` . the query is ```select count(*) from some_table where st_distance(resto_st_point, st_point(116.459717 , 39.955734, 1)) < 3000 ``` the table config is ``` "fieldConfigList": [ { "name": "resto_st_point", "encodingType": "RAW", "indexType": "H3", "properties": { "resolutions": "12" } } ], ..... "noDictionaryColumns": [ "resto_st_point" ],``` and if i change the threshold to 300 (meters), the index hits.
  @mayanks: @yupeng ^^
  @yupeng: check the query plan in trace mode and see why index is not in use
  @wcxzjtz: let me check.
  @wcxzjtz: is this the right trace: ``` "0_0": [ { "H3IndexFilterOperator Time": 0 }, { "RangeDocIdSetOperator Time": 0 }, { "ProjectionOperator Time": 0 }, { "RangeDocIdSetOperator Time": 0 }, { "ProjectionOperator Time": 0 },``` not sure how to interpret it.
  @wcxzjtz: and this is for meter is 300. ``` { "H3IndexFilterOperator Time": 6 }, { "DocIdSetOperator Time": 6 }, { "ProjectionOperator Time": 6 },```
  @wcxzjtz: cc @yupeng @mayanks
@wcxzjtz: another question related to geoindex, from the , seems only `ST_Distance` can take advantage of h3 index, how about `ST_Contains` ?
  @xiangfu0: @yupeng
  @yupeng: not yet. there's a PR on this but has been stale for a while
  @wcxzjtz: got it.
  @xiangfu0: Shall we revisit this PR?
  @wcxzjtz: yeah. we kinda need that. we did query on a table with about 4M records, it is pretty slow. take about 37s.
  @wcxzjtz: i can probably take that.
@tanmeshmishra97: @tanmeshmishra97 has joined the channel
@weixiang.sun: Does Pinot provide any tool to merge small segments into bigger segments? We have mis-configuration creating a lot of small segments. This is problematic. I am wondering if we can mitigate it by merging the small segments.
  @mark.needham: There is a tool -
  @mark.needham:

#random


@working1634: @working1634 has joined the channel
@tanmeshmishra97: @tanmeshmishra97 has joined the channel

#troubleshooting


@working1634: @working1634 has joined the channel
@4shivamsajwan: hi everyone, I am storing 0 as a value in a column but on pinot UI it is showing -- instead of 0. And when I am fetching the data using where value = 0 the rows are getting fetched. Any idea on this?
  @mayanks: How are you querying?
  @4shivamsajwan: lets say I have a column value1. so query will be like select * where value1 = 0
  @mayanks: Can you call the broker rest api and see what the response looks like?
  @mayanks:
@tanmeshmishra97: @tanmeshmishra97 has joined the channel
@tony: We are experimenting with *Ingestion Transformations*, specifically having a *filterConfig*. It works as expected, but there seems to be no way change the filterConfig. Running `pinot-admin.sh AddTable` with a changed *filterConfig* updates the table description shown in the UI, but does not reflect in the data actually ingested; we had to drop and recreate the table, which workable for testing but not later. Is there a way to get Pinot to actually use the new *filterConfig*?

#getting-started


@mark.needham: We've been writing up a bunch of recipes that cover some common Pinot use cases -
--------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@pinot.apache.org For additional commands, e-mail: dev-h...@pinot.apache.org

Reply via email to