v0.9.7

CSV v0.9.7

Diff since v0.9.6

Closed issues:

• Method ambiguity in _promote when promoting x::T to T if T <: PooledVector (#925)
• CSV v0.9.6 fails to precompile as solo package on fresh 1.6.1 project on Ubuntu (#927)
• CSV.Rows errors when types for a column is specified to be Symbol (#929)

Merged pull requests:

• Fix ambiguity of "promoting" a PooledVector to its same type (#926) (@nickrobinson251)
• Move promotions.csv file to src/ since we use it for precompile (#930) (@quinnj)
• Fix macro hygiene issue in CSV.Rows for custom types (#931) (@quinnj)

v0.9.6

CSV v0.9.6

Diff since v0.9.5

Merged pull requests:

• Bump version (#924) (@nickrobinson251)

v0.9.5

CSV v0.9.5

Diff since v0.9.4

Closed issues:

• MethodError: no method matching checkbounds(::CSV.PointerString, ::Int64) (#884)
• Vector of inputs enhancements (#891)
• Add example of pooling non-string columns (#895)
• Precompilation error due to WeakRefStrings (#912)
• Incorrect BigFloat read (#916)
• buffer_to_tempfile for Gzip stream - multiple tempfiles, one not cleared up (#917)
• out-of-bounds writes with @inbounds (#918)
• Unable to use CSV in Julia 1.6 (#919)
• Warnings when reading csv file (#921)

Merged pull requests:

• Add example of non-string column pooling in docs (#914) (@quinnj)
• Enhancements to vector of inputs to CSV.File (#920) (@quinnj)
• Ensure tempfile for compressed input is removed when uncompressed (#922) (@quinnj)
• Take dependency on InlineStrings package (#923) (@quinnj)

v0.9.4

CSV v0.9.4

Diff since v0.9.3

Closed issues:

• Race condition in 0.9 (#888)
• multithreaded reading of tsv gives TaskFailedException (#907)
• Reading from IO buffer: 0.8 vs. 0.9 (#908)

Merged pull requests:

• Try to ensure refpools are promoted between string/inlinestring (#909) (@quinnj)

v0.9.3

CSV v0.9.3

Diff since v0.9.2

Closed issues:

• Difference between File and Rows (#893)
• CodeUnits input broken with CSV.jl 0.9 (#894)
• Allow passing function for types/pool keyword args (#896)
• when informing data types, use functions to parse type (e.g. DateTime) (#899)
• Error while trying to writing DataFrames with >= 2^16 columns (#900)
• Problems with type conversions in CSV 0.9.x (#901)
• Columns with missing data are filled with stale data using CSV.Rows (#903)

Merged pull requests:

• Fix nonstandard type passed as single types and fix deprecations (#902) (@quinnj)
• Fix CSV.Rows when missing values are parsed (#904) (@quinnj)
• Support CodeUnits as valid input for CSV.File/CSV.Rows (#905) (@quinnj)
• Allow passing function for types/pool keyword arguments (#906) (@quinnj)

v0.9.2

CSV v0.9.2

Diff since v0.9.1

Closed issues:

• Race condition in 0.9 (#888)

Merged pull requests:

• Attempt to make reading from S3Paths work again (#892) (@quinnj)
• Declare isrowtable(::Rows) == true (#897) (@nickrobinson251)
• Refactor internal pooling mechanics (#898) (@quinnj)

v0.9.1

CSV v0.9.1

Diff since v0.9.0

Closed issues:

• re-exporting InlineString[X] types (#887)

Merged pull requests:

• Jq/docs (#886) (@quinnj)
• Export String[X] types from WeakRefStrings.jl (#889) (@bkamins)

v0.9.0

CSV v0.9.0

Diff since v0.8.5

Closed issues:

• Saving a DataFrame with many columns throws obscure error (#635)
• Bulk importing of CSV files (#712)
• CSV.File broken with timestamps in 0.7.7 (#729)
• Slow reading of gzipped files (#733)
• Exporting DataFrame to CSV also dumps array datatype (#740)
• Allow per-column missing value specification (#768)
• Specify column types from dictionary (#802)
• Automatic character type encoding? (#804)
• problem of CSV.read in Jupyter notebook (#806)
• Parsing error when setting column type to ZonedDateTime (#807)
• Maybe allow mixing string/symbol/int in select/drop options (#810)
• CSV.Rows with gzipped files running out of memory (#815)
• Segfault during CSV reading (#816)
• Hang with multithreaded reading (#817)
• Error for nothing in CSV.write should provide copy-pastable code (#818)
• CSV.File: kwargs missingstrings/missingstring (#821)
• MethodError: no method matching fuzzymatch(::Dict{Symbol, Int64}, ::QuoteNode) (#824)
• CSV.File seems to not handle pool kwarg correctly (#829)
• Additional Error Handling with multithreaded CSV file read (#830)
• How to use CSV.Rows() and CSV.Rows2? (#833)
• Reading Transposed CSV with unequal rows reads only minimum size of data (#834)
• quotechar = '\0' into documentation (#836)
• Feature Request: Custom parsing functions without overloading Base.parse (#838)
• SentinelArrays.SentinelCollisionError when reading file (#839)
• Missing method for SentinelArray (#842)
• How to use CSV.read() (#843)
• Reading String with zero byte fails. (#845)
• Providing a single type to types kwarg (#851)
• Can't turn off threading (#852)
• Default column types when reading CSV files (#854)
• Float with many decimal places unexpectedly read as 0.0 (#855)
• select and drop keywords seems to be broken (#856)
• Using select with CSV.Rows (#858)
• Error when file is not found (#859)
• Error when loading complex NaN's (#860)
• add AbstractVecOrMat method to CSV.write (#861)
• Cannot add CSV 0.8.5 and Parsers 2.0.2 into the same project (#867)
• Precompilation fails (#868)
• Keyword typemap has no effect (#871)
• Add writing of compressed CSV files (#874)
• Precompiling CSV - ArgumentError: Package CSV does not have Parsers in its dependencies (#878)
• Nested task error (#879)
• Can't read csv in Julia (#881)

Merged pull requests:

• make "Edit on Github" points to main branch (#835) (@Moelf)
• CSV parsing internals refactoring (#837) (@quinnj)
• Keyword argument cleanup in preparation for 1.0 release (#846) (@quinnj)
• Fix ByteValueCounter out of bounds indexing (#847) (@quinnj)
• Add new quoted keyword arg to allow disabling of checking for quoted (#848) (@quinnj)
• Use sch.names to get schema names instead of relying on type parameter (#849) (@quinnj)
• Update CSV to support Parsers 2.0 (#862) (@quinnj)
• Improve error message in case input file doesn't exist (#863) (@quinnj)
• Add automatic support for reading gzipped csv files (#864) (@quinnj)
• Allow turning off missing value parsing (#865) (@quinnj)
• Fix select/drop on CSV.Rows (#866) (@quinnj)
• Overhaul CSV.jl docs (#869) (@quinnj)
• Add extra validation on types/dateformat/pool keyword args (#870) (@quinnj)
• Respect typemap in type promotion cases while parsing (#872) (@quinnj)
• Allow passing vector of inputs to CSV.File/CSV.read (#873) (@quinnj)
• Improve time-to-first-read via precompile + despecialization (#875) (@quinnj)
• Support SubArray{UInt8, 1, Vector{UInt8}} as valid source (#876) (@quinnj)
• Re-organise docs a little to make navigation easier (#877) (@nickrobinson251)
• A little refactoring to improve inference and precompilation (#880) (@quinnj)
• Support gzip compression when writing csv output (#883) (@quinnj)
• Fix a few allocation/inference issues (#885) (@quinnj)

v0.8.5

CSV v0.8.5

Diff since v0.8.4

Closed issues:

• typemap doesn't work reliably in Juno (#797)
• Allow reading string columns as symbols (#812)
• CSV.read fail require sink argument (#814)
• Could the CSV.write argument order be interchanged ? (#819)
• UndefRefError when concatenating vector of dataframes (#820)
• Do not pool values by default? (#822)
• Formatting of bulleted items in docstrings (#825)
• Issue with Tables.rowtable (#827)
• Error when reading CSV files. (#832)

Merged pull requests:

• Don't make PointerString a subtype of AbstractString (#823) (@andreasnoack)
• tweak format of bulleted docstrings (#826) (@cormullion)
• Fix CSV.File schema for pooled columns when multithreaded parsing (#828) (@quinnj)

v0.8.4

CSV v0.8.4

Diff since v0.8.3

Closed issues:

• CSV reading memory usage (#794)
• read() slower than classical readdlm() (#795)
• footerskip Windows/Linux inconsistency (#796)
• TypeError: in new, expected CSV.PooledString, got a value of type String (#808)
• Issues with CSV.read in Atom (#809)

Merged pull requests:

• Scrub CSV.File types for Tables.Schema (#811) (@quinnj)
• Add tests for #796 (#813) (@quinnj)