Parquet.jl Versions Save

Parquet v0.8.4

Diff since v0.8.3

Closed issues:

• Reading Int64 columns throws Inexact Error (#149)
• Failing to precompile Snappy.jl..... (#153)
• KeyError when writing DataFrame to Parquet (#157)
• Not thread safe? (#160)
• Support [email protected] (#164)

Merged pull requests:

• Update writer.jl usage of Tables.Columns (#158) (@quinnj)
• Add compat for Snappy v0.4 (#166) (@giordano)
• Move CI to GitHub Actions (#167) (@giordano)

Parquet v0.8.3

Diff since v0.8.2

Closed issues:

• update to support Missings.jl version 1.0 (#147)

Merged pull requests:

• updated Missings.jl to v1 (#148) (@xiaodaigh)
• Update to 0.8.3 release to support latest categorical arrays. (#151) (@xiaodaigh)

Parquet v0.8.2

Diff since v0.8.1

Merged pull requests:

• Support CategoricalArrays 0.10 and bump version to 0.8.2 (#146) (@nalimilan)

Parquet v0.8.1

Diff since v0.8.0

Closed issues:

• Can't read a specific file (#52)
• Integration with Tables.jl (#70)
• Mention Tables.jl and possibly DataFrames in the docs (#98)
• add a high-level, easy to use read_parquet (#100)
• Add Finaliser to Parquet.File (#119)
• Write is very slow due to call to Base.summarysize (#126)
• How to read Parquet files from a structured folder? (#139)

Merged pull requests:

• ENH: allow writer to write to arbitrary IO (#127) (@sglyon)
• Attempt to improve efficiency of read_parquet by using ChainedVector … (#128) (@quinnj)
• use Table methods on read_parquet result (#130) (@tanmaykm)
• read_parquet improvements (#134) (@tanmaykm)
• avoid warning when mmap mode is set (#135) (@tanmaykm)
• finalizer for Parquet.File to close file handle (#136) (@tanmaykm)
• support reading metadata only files (#137) (@tanmaykm)
• support for reading partitioned datasets (#138) (@tanmaykm)
• Extrapolate size of sample for writing. (#140) (@altre)
• fix reading dataset without metadata file (#141) (@tanmaykm)
• fill missing partitioned col from partition path (#142) (@tanmaykm)
• bump version (#143) (@tanmaykm)

Parquet v0.8.0

Diff since v0.7.1

Closed issues:

• Out of range error (#120)

Merged pull requests:

• Simple reader (#116) (@xiaodaigh)
• handle trailing bytes in bitpacked run (#123) (@tanmaykm)
• bump version for tagging (#124) (@tanmaykm)

Parquet v0.7.1

Diff since v0.7.0

Closed issues:

• overloading of Base.eltype (#107)
• "UndefVarError: File not defined" when trying to use Parquet.File (#109)
• Inconsistent Results when Querying Rows (#111)
• update to Thrift 0.8 (#113)
• Update to MemPool v0.3 (#114)

Merged pull requests:

• [Tiny] Fix ColCursor length bug (#110) (@sa-)
• updated Categoricalarrays to 0.9 (#115) (@xiaodaigh)
• correct Base.eltype overloading (#117) (@tanmaykm)
• fix loading parquet file with a start offset (#118) (@tanmaykm)

Parquet v0.7.0

Diff since v0.6.1

Closed issues:

• Write parquet file (#1)
• String columns read back as Vector{UInt8} (#94)
• Why ParFile instead of ParquetFile? (#99)
• Corrupt reads with large data files (#105)

Merged pull requests:

• reader improvements (#97) (@tanmaykm)
• read decimals encoded as fixed width byte arrays (#101) (@tanmaykm)
• update to use Thrift v0.8 (#102) (@tanmaykm)
• improve tests and coverage (#103) (@tanmaykm)
• support some more logical type mappings (#104) (@tanmaykm)
• rename ParFile to Parquet.File (#106) (@tanmaykm)

Parquet v0.6.1

Diff since v0.6.0

Merged pull requests:

• fix setrow throw BoundsError (#95) (@ssikdar1)
• Accept any Tables.jl-compatible source to write_parquet (#96) (@quinnj)

Parquet v0.6.0

Diff since v0.5.2

Closed issues:

• Bug with BatchedColumnCursor reader (#89)

Merged pull requests:

• Parquet Writer (#66) (@xiaodaigh)
• Added a batched columns iterator (#83) (@tanmaykm)
• fix order of definition and repetition levels (#86) (@tanmaykm)
• read files with nested arrays (#87) (@tanmaykm)
• fix BatchedColumnsIterator missing value handling (#91) (@tanmaykm)
• bump version to v0.6 (#92) (@tanmaykm)
• Update Parquet.jl (#93) (@xiaodaigh)

Parquet v0.5.2

Diff since v0.5.1

Closed issues:

• values(....)[1] is outputting too many values for column with missing? (#81)
• string columns with missing are read back at Int32 (#82)

Merged pull requests:

• fix dict page mapping for missing values (#84) (@tanmaykm)