Kimahriman commented on code in PR #915:
URL: https://github.com/apache/datafusion-comet/pull/915#discussion_r1759619909
##########
spark/src/main/scala/org/apache/spark/sql/comet/CometScanExec.scala:
##########
@@ -141,8 +141,30 @@ case class CometScanExec(
if (wrapped == null) Map.empty else wrapped.metadata
override def verboseStringWithOperatorId(): String = {
- getTagValue(QueryPlan.OP_ID_TAG).foreach(id =>
wrapped.setTagValue(QueryPlan.OP_ID_TAG, id))
- wrapped.verboseStringWithOperatorId()
+ val metadataStr = metadata.toSeq.sorted
Review Comment:
Just removing setting the `OP_ID_TAG` leads to a formatted explain of:
```
== Physical Plan ==
* ColumnarToRow (2)
+- CometScan parquet (1)
(unknown) Scan parquet
Output [20]: [_1#0, _2#1, _3#2, _4#3, _5#4L, _6#5, _7#6, _8#7, _9#8, _10#9,
_11#10L, _12#11, _13#12, _14#13, _15#14, _16#15, _17#16, _18#17, _19#18, _20#19]
Batched: true
Location: InMemoryFileIndex
[file:/Users/abinford/projects/arrow-datafusion-comet/spark/target/tmp/spark-c0b82b7c-3de1-431b-96bf-56cb37f3a463/test.parquet]
ReadSchema:
struct<_1:boolean,_2:tinyint,_3:smallint,_4:int,_5:bigint,_6:float,_7:double,_8:string,_9:smallint,_10:int,_11:bigint,_12:decimal(20,0),_13:string,_14:binary,_15:decimal(5,2),_16:decimal(18,10),_17:decimal(38,37),_18:timestamp,_19:timestamp,_20:date>
(2) ColumnarToRow [codegen id : 1]
Input [20]: [_1#0, _2#1, _3#2, _4#3, _5#4L, _6#5, _7#6, _8#7, _9#8, _10#9,
_11#10L, _12#11, _13#12, _14#13, _15#14, _16#15, _17#16, _18#17, _19#18, _20#19]
```
which is the whole reason it was added in the first place, to fix the
unknown operator ID
https://github.com/apache/datafusion-comet/commit/bc35fa5a415f4a137177ab4fc66bd0167f9f23d3
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
