[jira] [Created] (ARROW-16983) Delta byte array encoder broken due to memory leak
Matt DePero created ARROW-16983: --- Summary: Delta byte array encoder broken due to memory leak Key: ARROW-16983 URL: https://issues.apache.org/jira/browse/ARROW-16983 Project: Apache Arrow Issue Type: Bug Components: Go, Parquet Reporter: Matt DePero The `DeltaByteArrayEncoder` has a memory leak due to a bug in how `EstimatedDataEncodedSize` is calculated. DeltaByteArrayEncoder extends `encoder` which calculates EstimatedDataEncodedSize by calling `Len()` on its `PooledBufferWriter` sink. DeltaByteArrayEncoder however does not write data to sink, instead writing data to `prefixEncoder` and `suffixEncoder` causing EstimatedDataEncodedSize to always return zero. -- This message was sent by Atlassian Jira (v8.20.10#820010)
[jira] [Created] (ARROW-16813) [Go][Parquet] Disabling dictionary encoding per column in writer config broken
Matt DePero created ARROW-16813: --- Summary: [Go][Parquet] Disabling dictionary encoding per column in writer config broken Key: ARROW-16813 URL: https://issues.apache.org/jira/browse/ARROW-16813 Project: Apache Arrow Issue Type: Bug Components: Go Reporter: Matt DePero Small bug in how we set column level dictionary encoding config, always set to true rather than respecting input value. -- This message was sent by Atlassian Jira (v8.20.7#820007)
[jira] [Created] (ARROW-16790) [Go][Parquet] Avoid allocating new memory when skipping rows
Matt DePero created ARROW-16790: --- Summary: [Go][Parquet] Avoid allocating new memory when skipping rows Key: ARROW-16790 URL: https://issues.apache.org/jira/browse/ARROW-16790 Project: Apache Arrow Issue Type: Improvement Components: Go Reporter: Matt DePero As referenced in [apache/arrow#13277|[https://github.com/apache/arrow/pull/13277],] we allocate scratch space memory every time we want to skip rows in the Go parquet reader despite this memory simply being discarded. We ideally should reuse scratch space when skipping to avoid these unnecessary allocations. -- This message was sent by Atlassian Jira (v8.20.7#820007)
[jira] [Reopened] (ARROW-16638) [Go][Parquet] Boolean column reader fails to skip rows
[ https://issues.apache.org/jira/browse/ARROW-16638?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Matt DePero reopened ARROW-16638: - Boolean skip still has issue when skipping a large number of rows, due to the buffer passed for `values` not being large enough. If batch size to skip is 1024, the buffer allocated for values when skipping a boolean column is only 128, resulting in an index out of bounds panic. > [Go][Parquet] Boolean column reader fails to skip rows > -- > > Key: ARROW-16638 > URL: https://issues.apache.org/jira/browse/ARROW-16638 > Project: Apache Arrow > Issue Type: Bug > Components: Go >Reporter: Matt DePero >Priority: Major > Labels: pull-request-available > Fix For: 9.0.0 > > Time Spent: 1.5h > Remaining Estimate: 0h > > Skipping values in the go parquet column reader is effectively implemented by > reading the target number of rows into scratch space which is then discarded. > In the boolean case, > [BytesRequired|https://github.com/apache/arrow/blob/4c21fd12f93e4853c03c05919ffb22c6bb8f09b0/go/parquet/file/column_reader.go#L439] > returns returns a scratch buffer that allocates one bit per row, however > that [same scratch > space|https://github.com/apache/arrow/blob/4c21fd12f93e4853c03c05919ffb22c6bb8f09b0/go/parquet/file/column_reader_types.gen.go#L212-L213] > is also attempted to be used for `defLvls` and `repLvls` (both int16), which > requires two bytes per row. Since the boolean `values` buffer is not large > enough to hold the same number of rows worth of def and rep levels, skipping > too many rows results in an index out of bounds panic. > > Note that for other column types, this does not seem to be an issue since the > buffer needed for `values` is always larger than the buffer needed for def > and rep levels, however there still seems to be no reason to include any > non-nil value to `cr.ReadBatch(...)` for [rep and def > lvls|https://github.com/apache/arrow/blob/4c21fd12f93e4853c03c05919ffb22c6bb8f09b0/go/parquet/file/column_reader_types.gen.go#L212-L213] > when skipping any column in the reader. -- This message was sent by Atlassian Jira (v8.20.7#820007)
[jira] [Created] (ARROW-16638) [Go][Parquet] Boolean column reader fails to skip rows
Matt DePero created ARROW-16638: --- Summary: [Go][Parquet] Boolean column reader fails to skip rows Key: ARROW-16638 URL: https://issues.apache.org/jira/browse/ARROW-16638 Project: Apache Arrow Issue Type: Bug Components: Go Reporter: Matt DePero Fix For: 9.0.0 Skipping values in the go parquet column reader is effectively implemented by reading the target number of rows into scratch space which is then discarded. In the boolean case, [BytesRequired|https://github.com/apache/arrow/blob/4c21fd12f93e4853c03c05919ffb22c6bb8f09b0/go/parquet/file/column_reader.go#L439] returns returns a scratch buffer that allocates one bit per row, however that [same scratch space|https://github.com/apache/arrow/blob/4c21fd12f93e4853c03c05919ffb22c6bb8f09b0/go/parquet/file/column_reader_types.gen.go#L212-L213] is also attempted to be used for `defLvls` and `repLvls` (both int16), which requires two bytes per row. Since the boolean `values` buffer is not large enough to hold the same number of rows worth of def and rep levels, skipping too many rows results in an index out of bounds panic. Note that for other column types, this does not seem to be an issue since the buffer needed for `values` is always larger than the buffer needed for def and rep levels, however there still seems to be no reason to include any non-nil value to `cr.ReadBatch(...)` for [rep and def lvls|https://github.com/apache/arrow/blob/4c21fd12f93e4853c03c05919ffb22c6bb8f09b0/go/parquet/file/column_reader_types.gen.go#L212-L213] when skipping any column in the reader. -- This message was sent by Atlassian Jira (v8.20.7#820007)
[jira] [Created] (ARROW-16563) [Go][Parquet] Plain boolean decoder is broken
Matt DePero created ARROW-16563: --- Summary: [Go][Parquet] Plain boolean decoder is broken Key: ARROW-16563 URL: https://issues.apache.org/jira/browse/ARROW-16563 Project: Apache Arrow Issue Type: Bug Components: Go Reporter: Matt DePero When reading parquet files with `PLAIN` encoded boolean values, `bitOffset` in the boolean decoder is not properly managed which results in garbled values being returned. -- This message was sent by Atlassian Jira (v8.20.7#820007)