This repository has been archived by the owner on Sep 18, 2023. It is now read-only.
[NSE-1171] Support merge parquet schema and read missing schema #1175
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
This PR could be tested in |
PHILO-HE
reviewed
Dec 5, 2022
...ard/src/main/scala/com/intel/oap/spark/sql/execution/datasources/arrow/ArrowFileFormat.scala
Outdated
Show resolved
Hide resolved
|
|
||
| val nullVectors = if (hashMissingColumns) { | ||
| val vectors = | ||
| ArrowWritableColumnVector.allocateColumns(batchSize, requiredSchema) |
There was a problem hiding this comment.
It looks these columnar vectors are allocated based on all required schema. And then keep truly null vectors for missing columns in ArrowUtils.scala. Can this part be optimized? It looks unnecessary to create null vectors for non-null column. Thanks!
There was a problem hiding this comment.
Hm, good catch, I would have a try.
...ndard/src/main/scala/com/intel/oap/spark/sql/execution/datasources/v2/arrow/ArrowUtils.scala
Outdated
Show resolved
Hide resolved
PHILO-HE
reviewed
Dec 6, 2022
| .getOrElse { | ||
| // The missing column need to be find in nullVectors | ||
| val nullVector = | ||
| nullVectors.find(_.getValueVector.getName.equalsIgnoreCase(field.name)).get |
There was a problem hiding this comment.
Just a small suggestion for code refinement. I think the different handling for case sensitive/insensitive can be simplified. We can define a lambda expression as below. Thus, we can directly use this eql in finding matched field, like anArray.find(x => eql(x, "MATCH_TARGET")), without separating the code logic with if/else. We can do the similar code refinement in other places. Right?
val eql = if (caseSensitive) {
(a: String, b: String) => a.equals(b)
} else {
(a: String, b: String) => a.equalsIgnoreCase(b)
}
jackylee-ch
changed the title
|
@jackylee-ch could you please also add a small Scala unit test for this feature? |
Sure |
zhouyuan
pushed a commit
to zhouyuan/native-sql-engine
that referenced
this pull request
Dec 14, 2022
…project#1175) * Support merge parquet schema and read missing schema * fix error * optimize null vectors * optimize code * optimize code * change code * add schema merge suite tests * add test for struct type
zhouyuan
added a commit
that referenced
this pull request
Dec 14, 2022
* [NSE-1170] Set correct row number in batch scan w/ partition columns (#1172) * [NSE-1171] Throw RuntimeException when reading duplicate fields in case-insensitive mode (#1173) * throw exception if one more columns matched in case insensitive mode * add schema check in arrow v2 * bump h2/pgsql version (#1176) * bump h2/pgsql version Signed-off-by: Yuan Zhou <yuan.zhou@intel.com> * ignore one failed test Signed-off-by: Yuan Zhou <yuan.zhou@intel.com> Signed-off-by: Yuan Zhou <yuan.zhou@intel.com> * [NSE-956] allow to write parquet with compression (#1014) This patch adds support for writing parquet with compression df.coalesce(1).write.format("arrow").option("parquet.compression","zstd").save(path) Signed-off-by: Yuan Zhou yuan.zhou@intel.com * [NSE-1161] Support read-write parquet conversion to read-write arrow (#1162) * add ArrowConvertExtension * do not convert parquet fileformat while writing to partitioned/bucketed/sorted output * fix cache failed * care about write codec * disable convertor extension by default * add some comments * remove wrong compress type check (#1178) Since the compresssion has been supported in #1014 . The extra compression check in ArrowConvertorExtension can be remove now. * fix to use right arrow branch (#1179) fix to use right arrow branch Signed-off-by: Yuan Zhou <yuan.zhou@intel.com> * [NSE-1171] Support merge parquet schema and read missing schema (#1175) * Support merge parquet schema and read missing schema * fix error * optimize null vectors * optimize code * optimize code * change code * add schema merge suite tests * add test for struct type * to use 1.5 branch arrow Signed-off-by: Yuan Zhou <yuan.zhou@intel.com> Signed-off-by: Yuan Zhou <yuan.zhou@intel.com> Signed-off-by: Yuan Zhou yuan.zhou@intel.com Co-authored-by: Jacky Lee <lijunqing@baidu.com>
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
What changes were proposed in this pull request?
This pr is trying to support Parquet Schema merge in ArrowFileFormat.infer_schema and support dealing with missing column or filter in Parquet reading.
How was this patch tested?
unit tests.