[GitHub] [arrow-site] wesm commented on a change in pull request #19: ARROW-6419: [Website] Blog post about Parquet C++ read performance improvements in Arrow 0.15
wesm commented on a change in pull request #19: ARROW-6419: [Website] Blog post about Parquet C++ read performance improvements in Arrow 0.15 URL: https://github.com/apache/arrow-site/pull/19#discussion_r321321708 ## File path: _posts/2019-09-03-faster-strings-cpp-parquet.md ## @@ -0,0 +1,238 @@ +--- +layout: post +title: "Faster C++ Apache Parquet performance on dictionary-encoded string data coming in Apache Arrow 0.15" +date: "2019-09-05 00:00:00 -0600" +author: Wes McKinney +categories: [application] +--- + + +We have been implementing a series of optimizations in the Apache Parquet C++ +internals to improve read and write efficiency (both performance and memory +use) for Arrow columnar binary and string data, with new "native" support for +Arrow's dictionary types. This should have a big impact on users of the C++, +MATLAB, Python, R, and Ruby interfaces to Parquet files. + +This post reviews work that was done and shows benchmarks comparing Arrow +0.12.1 with the current development version (to be released soon as Arrow +0.15.0). + +# Summary of work + +One of the largest and most complex optimizations involves encoding and +decoding Parquet files' internal dictionary-encoded data streams to and from +Arrow's in-memory dictionary-encoded `DictionaryArray` +representation. Dictionary encoding is a compression strategy in Parquet, and +there is no formal "dictionary" or "categorical" type. I will go into more +detail about this below. + +Some of the particular JIRA issues related to this work include: + +- Vectorize comparators for computing statistics ([PARQUET-1523][1]) +- Read binary directly data directly into DictionaryBuilder + ([ARROW-3769][2]) +- Writing Parquet's dictionary indices directly into DictionaryBuilder + ([ARROW-3772][3]) +- Write dense (non-dictionary) Arrow arrays directly into Parquet data encoders + ([ARROW-6152][4]) +- Direct writing of arrow::DictionaryArray to Parquet column writers ([ARROW-3246][5]) +- Supporting changing dictionaries ([ARROW-3144][6]) +- Internal IO optimizations and improved raw `BYTE_ARRAY` encoding performance + ([ARROW-4398][7]) + +One of the challenges of developing the Parquet C++ library is that we +maintain low-level read and write APIs that do not involve the Arrow columnar +data structures. So we have had to take care to do Arrow-related optimizations +without impacting non-Arrow Parquet users, which includes database systems like +Clickhouse and Vertica. + +# Background: how Parquet files do dictionary encoding + +Many direct and indirect users of Apache Arrow use dictionary encoding to +improve performance and memory use on binary or string data types that include +many repeated values. pandas users will know this as the [Categorical type][8] +while in R such encoding is known as [`factor`][9]. In the Arrow C++ library +and various bindings we have the `DictionaryArray` object for representing such +data in memory. + +For example, an array such as + +``` +['apple', 'orange', 'apple', NULL, 'orange', 'orange'] +``` + +has dictionary-encoded form + +``` +dictionary: ['apple', 'orange'] +indices: [0, 1, 0, NULL, 1, 1] +``` + +The [Parquet format uses dictionary encoding][10] to compress data, and it is +used for all Parquet data types, not just binary or string data. Parquet +further uses bit-packing and run-length encoding (RLE) to compress the +dictionary indices, so if you had data like + +``` +['apple', 'apple', 'apple', 'apple', 'apple', 'apple', 'orange'] +``` + +the indices would be encoded like + +``` +[rle-run=(6, 0), + bit-packed-run=[1]] +``` + +The full details of the rle-bitpacking encoding are found in the [Parquet +specification][10]. + +When writing a Parquet file, most implementations will use dictionary encoding +to compress a column until the dictionary itself reaches a certain size +threshold, usually around 1 megabyte. At this point, the column writer will +"fall back" to `PLAIN` encoding where values are written end-to-end in "data +pages" and then usually compressed with Snappy or Gzip. See the following rough +diagram: + + + + + +# Faster reading and writing of dictionary-encoded data + +When reading a Parquet file, the dictionary-encoded portions are usually +materialized to their non-dictionary-encoded form, causing binary or string +values to be duplicated in memory. So an obvious (but not trivial) optimization +is to skip this "dense" materialization. There are several issues to deal with: + +* A Parquet file often contains multiple ColumnChunks for each semantic column, + and the dictionary values may be different in each ColumnChunk +* We must gracefully handle the "fall back" portion which is not + dictionary-encoded + +We pursued several avenues to help with this: + +* Allowing each `arrow::DictionaryArray` to have a different dictionary + (before, the dictionary was part of the `DictionaryType`, which caused + problems) +* We enabled the Parquet dictionary indices to be directly
[GitHub] [arrow-site] wesm commented on a change in pull request #19: ARROW-6419: [Website] Blog post about Parquet C++ read performance improvements in Arrow 0.15
wesm commented on a change in pull request #19: ARROW-6419: [Website] Blog post about Parquet C++ read performance improvements in Arrow 0.15 URL: https://github.com/apache/arrow-site/pull/19#discussion_r320453511 ## File path: _posts/2019-09-03-faster-strings-cpp-parquet.md ## @@ -0,0 +1,233 @@ +--- +layout: post +title: "Faster C++ Apache Parquet performance on string-heavy data coming in Apache Arrow 0.15" +date: "2019-09-01 00:00:00 -0600" +author: Wes McKinney +categories: [application] +--- + + +We have been implementing a series of optimizations in the Apache Parquet C++ +internals to improve read and write efficiency (both performance and memory +use) for Arrow columnar binary and string types, including native support for +Arrow's dictionary types. This should have a big impact on users of the C++, +MATLAB, Python, R, and Ruby interfaces to Parquet files. + +We discuss the work that was done and show benchmarks comparing Arrow 0.11.0 +(released in October, 2018) with the current development version (to be +released soon as Arrow 0.15.0). + +# Summary of work + +One of the challenges of developing the Parquet C++ library is that we +maintain low-level read and write APIs that do not involve the Arrow columnar +data structures. So we have had to take care to do Arrow-related optimizations +without impacting non-Arrow Parquet users, which includes database systems like +Clickhouse and Vertica. + +One of the largest and most complex optimizations involves encoding and +decoding Parquet files' internal dictionary-encoded data streams to and from +Arrow's in-memory dictionary-encoded `DictionaryArray` +representation. Dictionary encoding is a compression strategy in Parquet, and +there is no formal "dictionary" or "categorical" type. I will go into more +detail about this below. + +Some of the particular JIRA issues related to this work include: + +- Vectorize comparators for computing statistics ([PARQUET-1523][1]) +- Read binary directly data directly into DictionaryBuilder + ([ARROW-3769][2]) +- Writing Parquet's dictionary indices directly into DictionaryBuilder + ([ARROW-3772][3]) +- Write dense (non-dictionary) Arrow arrays directly into Parquet data encoders + ([ARROW-6152][4]) +- Direct writing of arrow::DictionaryArray to Parquet column writers ([ARROW-3246][5]) +- Supporting changing dictionaries ([ARROW-3144][6]) +- Internal IO optimizations and improved raw `BYTE_ARRAY` encoding performance + ([ARROW-4398][7]) + +# Background: how Parquet files do dictionary encoding + +Many direct and indirect users of Apache Arrow use dictionary encoding to +improve performance and memory use on binary or string data types that include +many repeated values. pandas users will know this as the [Categorical type][8] Review comment: Lowercase is the official styling of the project name =) 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. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [arrow-site] wesm commented on a change in pull request #19: ARROW-6419: [Website] Blog post about Parquet C++ read performance improvements in Arrow 0.15
wesm commented on a change in pull request #19: ARROW-6419: [Website] Blog post about Parquet C++ read performance improvements in Arrow 0.15 URL: https://github.com/apache/arrow-site/pull/19#discussion_r320453000 ## File path: _posts/2019-09-03-faster-strings-cpp-parquet.md ## @@ -0,0 +1,233 @@ +--- +layout: post +title: "Faster C++ Apache Parquet performance on string-heavy data coming in Apache Arrow 0.15" +date: "2019-09-01 00:00:00 -0600" +author: Wes McKinney +categories: [application] +--- + + +We have been implementing a series of optimizations in the Apache Parquet C++ +internals to improve read and write efficiency (both performance and memory +use) for Arrow columnar binary and string types, including native support for +Arrow's dictionary types. This should have a big impact on users of the C++, +MATLAB, Python, R, and Ruby interfaces to Parquet files. + +We discuss the work that was done and show benchmarks comparing Arrow 0.11.0 +(released in October, 2018) with the current development version (to be +released soon as Arrow 0.15.0). + +# Summary of work + +One of the challenges of developing the Parquet C++ library is that we +maintain low-level read and write APIs that do not involve the Arrow columnar +data structures. So we have had to take care to do Arrow-related optimizations +without impacting non-Arrow Parquet users, which includes database systems like +Clickhouse and Vertica. + +One of the largest and most complex optimizations involves encoding and +decoding Parquet files' internal dictionary-encoded data streams to and from +Arrow's in-memory dictionary-encoded `DictionaryArray` +representation. Dictionary encoding is a compression strategy in Parquet, and +there is no formal "dictionary" or "categorical" type. I will go into more +detail about this below. + +Some of the particular JIRA issues related to this work include: + +- Vectorize comparators for computing statistics ([PARQUET-1523][1]) +- Read binary directly data directly into DictionaryBuilder + ([ARROW-3769][2]) +- Writing Parquet's dictionary indices directly into DictionaryBuilder + ([ARROW-3772][3]) +- Write dense (non-dictionary) Arrow arrays directly into Parquet data encoders + ([ARROW-6152][4]) +- Direct writing of arrow::DictionaryArray to Parquet column writers ([ARROW-3246][5]) +- Supporting changing dictionaries ([ARROW-3144][6]) +- Internal IO optimizations and improved raw `BYTE_ARRAY` encoding performance + ([ARROW-4398][7]) + +# Background: how Parquet files do dictionary encoding + +Many direct and indirect users of Apache Arrow use dictionary encoding to +improve performance and memory use on binary or string data types that include +many repeated values. pandas users will know this as the [Categorical type][8] +while in R such encoding is known as [`factor`][9]. In the Arrow C++ library +and various bindings we have the `DictionaryArray` object for representing such +data in memory. + +For example, an array such as + +``` +['apple', 'orange', 'apple', NULL, 'orange', 'orange'] +``` + +has dictionary-encoded form + +``` +dictionary: ['apple', 'orange'] +indices: [0, 1, 0, NULL, 1, 1] +``` + +The [Parquet format uses dictionary encoding][10] to compress data, and it is +used for all Parquet data types, not just binary or string data. Parquet +further uses bit-packing and run-length encoding (RLE) to compress the +dictionary indices, so if you had data like + +``` +['apple', 'apple', 'apple', 'apple', 'apple', 'apple', 'orange'] +``` + +the indices would be encoded like + +``` +[rle-run=(6, 0), + bit-packed-run=[1]] +``` + +The full details of the rle-bitpacking encoding are found in the Parquet +specification. + +When writing a Parquet file, most implementations will use dictionary encoding +to compress a column until the dictionary itself reaches a certain size +threshold, usually around 1 megabyte. At this point, the column writer will +"fall back" to `PLAIN` encoding where values are written end-to-end in "data +pages" and then usually compressed with Snappy or Gzip. See the following rough +diagram: + + + + + +When reading a Parquet file, the dictionary-encoded portions are usually +materialized to their non-dictionary-encoded form, causing binary or string +values to be duplicated in memory. So an obvious (but not trivial) optimization +is to skip this "dense" materialization. There are several issues to deal with: + +* A Parquet file often contains multiple ColumnChunks for each semantic column, + and the dictionary values may be different in each ColumnChunk +* We must gracefully handle the "fall back" portion which is not + dictionary-encoded + +We pursued several avenues to help with this: + +* Allowing each `arrow::DictionaryArray` to have a different dictionary + (before, the dictionary was part of the `DictionaryType`, which caused + problems) +* We enabled the Parquet dictionary indices to be directly written into an + Arrow `DictionaryBuilder` without
