feat: add stats to convert-to-delta operation #2491
There are no files selected for viewing
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 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
| Original file line number | Diff line number | Diff line change |
|---|---|---|
|
|
@@ -4,6 +4,7 @@ use std::time::{SystemTime, UNIX_EPOCH}; | |
| use std::{collections::HashMap, ops::AddAssign}; | ||
|
|
||
| use indexmap::IndexMap; | ||
| use parquet::file::metadata::ParquetMetaData; | ||
| use parquet::format::FileMetaData; | ||
| use parquet::schema::types::{ColumnDescriptor, SchemaDescriptor}; | ||
| use parquet::{basic::LogicalType, errors::ParquetError}; | ||
|
|
@@ -66,6 +67,32 @@ pub fn create_add( | |
| }) | ||
| } | ||
|
|
||
| // As opposed to `stats_from_file_metadata` which operates on `parquet::format::FileMetaData`, | ||
| // this function produces the stats by reading the metadata from already written out files. | ||
| // | ||
| // Note that the file metadata used here is actually `parquet::file::metadata::FileMetaData` | ||
| // which is a thrift decoding of the `parquet::format::FileMetaData` which is typically obtained | ||
| // when flushing the write. | ||
| pub(crate) fn stats_from_parquet_metadata( | ||
| partition_values: &IndexMap<String, Scalar>, | ||
| parquet_metadata: &ParquetMetaData, | ||
| num_indexed_cols: i32, | ||
| stats_columns: &Option<Vec<String>>, | ||
| ) -> Result<Stats, DeltaWriterError> { | ||
| let num_rows = parquet_metadata.file_metadata().num_rows(); | ||
| let schema_descriptor = parquet_metadata.file_metadata().schema_descr_ptr(); | ||
| let row_group_metadata = parquet_metadata.row_groups().to_vec(); | ||
|
|
||
| stats_from_metadata( | ||
| partition_values, | ||
| schema_descriptor, | ||
| row_group_metadata, | ||
| num_rows, | ||
| num_indexed_cols, | ||
| stats_columns, | ||
| ) | ||
| } | ||
|
|
||
| fn stats_from_file_metadata( | ||
| partition_values: &IndexMap<String, Scalar>, | ||
| file_metadata: &FileMetaData, | ||
|
|
@@ -75,27 +102,46 @@ fn stats_from_file_metadata( | |
| let type_ptr = parquet::schema::types::from_thrift(file_metadata.schema.as_slice()); | ||
| let schema_descriptor = type_ptr.map(|type_| Arc::new(SchemaDescriptor::new(type_)))?; | ||
|
|
||
| let row_group_metadata: Vec<RowGroupMetaData> = file_metadata | ||
| .row_groups | ||
| .iter() | ||
| .map(|rg| RowGroupMetaData::from_thrift(schema_descriptor.clone(), rg.clone())) | ||
| .collect::<Result<Vec<RowGroupMetaData>, ParquetError>>()?; | ||
|
|
||
| stats_from_metadata( | ||
| partition_values, | ||
| schema_descriptor, | ||
| row_group_metadata, | ||
| file_metadata.num_rows, | ||
| num_indexed_cols, | ||
| stats_columns, | ||
| ) | ||
| } | ||
|
|
||
| fn stats_from_metadata( | ||
| partition_values: &IndexMap<String, Scalar>, | ||
| schema_descriptor: Arc<SchemaDescriptor>, | ||
| row_group_metadata: Vec<RowGroupMetaData>, | ||
| num_rows: i64, | ||
| num_indexed_cols: i32, | ||
| stats_columns: &Option<Vec<String>>, | ||
| ) -> Result<Stats, DeltaWriterError> { | ||
| let mut min_values: HashMap<String, ColumnValueStat> = HashMap::new(); | ||
| let mut max_values: HashMap<String, ColumnValueStat> = HashMap::new(); | ||
| let mut null_count: HashMap<String, ColumnCountStat> = HashMap::new(); | ||
|
|
||
| let idx_to_iterate = if let Some(stats_cols) = stats_columns { | ||
| schema_descriptor | ||
| .columns() | ||
| .into_iter() | ||
| .enumerate() | ||
| .filter_map(|(index, col)| { | ||
| if stats_cols.contains(&col.name().to_string()) { | ||
| Some(index) | ||
| } else { | ||
| None | ||
| } | ||
| }) | ||
| .collect() | ||
| } else if num_indexed_cols == -1 { | ||
| (0..schema_descriptor.num_columns()).collect::<Vec<_>>() | ||
|
|
@@ -149,7 +195,7 @@ fn stats_from_file_metadata( | |
| Ok(Stats { | ||
| min_values, | ||
| max_values, | ||
| num_records: num_rows, | ||
| null_count, | ||
| }) | ||
| } | ||
|
|
@@ -262,18 +308,8 @@ impl StatsScalar { | |
| v.max_bytes() | ||
| }; | ||
|
|
||
| let val = if val.len() <= 16 { | ||
| i128::from_be_bytes(sign_extend_be(val)) as f64 | ||
|
Comment on lines
-265
to
+312
There was a problem hiding this comment. Looks like this was causing those test failures. In particular note that the above would wrongly extend slices that are not a power of 2 long. This is the case that occured for the failing tests, since this: delta-rs/python/tests/conftest.py Line 218 in 81593e9 would result in min [0, 39, 16] and max [0, 54, 176] fixed length byte arrays, and these would in turn be coerced to [0, 39, 16, 0] an [0, 54, 176, 0], instead of [0, 0, 39, 16] and [0, 0, 54, 176] respectively. Contrast with arrow-rs/datafusion where the extension happens at the beginning of the array: Granted I'm not sure what/how encodes the min value "10.000" to [0, 39, 16] in the first place, aside from that it occurs in pyarrow. There was a problem hiding this comment. Hi @ion-elgreco, can you approve the rest of the workflows? |
||
| } else { | ||
| return Err(DeltaWriterError::StatsParsingFailed { | ||
| debug_value: format!("{val:?}"), | ||
|
|
@@ -315,6 +351,19 @@ impl StatsScalar { | |
| } | ||
| } | ||
|
|
||
| /// Performs big endian sign extension | ||
| /// Copied from arrow-rs repo/parquet crate: | ||
| /// https://github.com/apache/arrow-rs/blob/b25c441745602c9967b1e3cc4a28bc469cfb1311/parquet/src/arrow/buffer/bit_util.rs#L54 | ||
| pub fn sign_extend_be<const N: usize>(b: &[u8]) -> [u8; N] { | ||
| assert!(b.len() <= N, "Array too large, expected less than {N}"); | ||
| let is_negative = (b[0] & 128u8) == 128u8; | ||
| let mut result = if is_negative { [255u8; N] } else { [0u8; N] }; | ||
| for (d, s) in result.iter_mut().skip(N - b.len()).zip(b) { | ||
| *d = *s; | ||
| } | ||
| result | ||
| } | ||
|
|
||
| impl From<StatsScalar> for serde_json::Value { | ||
| fn from(scalar: StatsScalar) -> Self { | ||
| match scalar { | ||
|
|
@@ -653,6 +702,17 @@ mod tests { | |
| }), | ||
| Value::from(1243124142314.423), | ||
| ), | ||
| ( | ||
| simple_parquet_stat!( | ||
| Statistics::FixedLenByteArray, | ||
| FixedLenByteArray::from(vec![0, 39, 16]) | ||
| ), | ||
| Some(LogicalType::Decimal { | ||
| scale: 3, | ||
| precision: 5, | ||
| }), | ||
| Value::from(10.0), | ||
| ), | ||
| ( | ||
| simple_parquet_stat!( | ||
| Statistics::FixedLenByteArray, | ||
|
|
||
Oops, something went wrong.
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.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This shouldn't be -1. You should use the same function I recently added in the write the get the configuration values:
get_num_idx_cols_and_stats_columnsThere was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Good catch, revised now.