fix(spark)!: preserve partition values in file scans - #1270
Draft
bvolpato wants to merge 1 commit into
Draft
Conversation
A partitioned Parquet read containing rows (1,10) and (2,20) round-trips as (1,NULL) and (2,NULL). The exporter lists leaf files with the full table schema, while the importer treats virtual partition columns as physical file columns. Export each selected partition as a scan of its data schema with typed partition literals, preserve the merged column order through emit, and combine partitions with UNION_ALL. This uses Spark's resolved values rather than reconstructing them from paths, including nulls, explicit types, overlapping columns, pruned indexes, and multiple roots. Serialize canonical file URIs and decode them as URIs so escaped partition values address the original files. Local paths with unescaped spaces remain accepted. The representation uses standard Substrait v0.102.0 relations and adds one scan/project branch per nonempty partition. BREAKING CHANGE: Spark 3.4 conversion now rejects partitioned file scans whose file and partition column names differ only by case under case-insensitive resolution. Resolve the schema collision before conversion; this prevents changing the source plan's values to the different partition values.
bvolpato
force-pushed
the
bvolpato/fix-spark-partition-values
branch
from
September 4, 2026 16:32
560db29 to
b93bd7e
Compare
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
A partitioned Parquet read containing rows (1,10) and (2,20) round-trips as (1,NULL) and (2,NULL). The exporter lists leaf files with the full table schema, while the importer treats virtual partition columns as physical file columns.
Export each selected partition as a scan of its data schema with typed partition literals, preserve the merged column order through emit, and combine partitions with UNION_ALL. This uses Spark's resolved values rather than reconstructing them from paths, including nulls, explicit types, overlapping columns, pruned indexes, and multiple roots.
Serialize canonical file URIs and decode them as URIs so escaped partition values address the original files. Local paths with unescaped spaces remain accepted. The representation uses standard Substrait v0.102.0 relations and adds one scan/project branch per nonempty partition.
BREAKING CHANGE: Spark 3.4 conversion now rejects partitioned file scans whose file and partition column names differ only by case under case-insensitive resolution. Resolve the schema collision before conversion; this prevents changing the source plan's values to the different partition values.