`zstd error` when trying to read file produced by Polars

[ForceBru]

Polars Python code

import polars as pl

df = pl.DataFrame({
    'DateTime': [
        '2018-02-01 00:00:00', '2018-02-02 00:00:00',
        '2018-02-03 00:00:00', '2018-02-04 00:00:00',
        '2018-02-05 00:00:00'
    ],
    'String': ['a', 'b', 'c', 'd', 'e'],
    'Float': [1.2, 3.4, 5.6, 7.8, 9.0],
}).select(
    pl.col('DateTime').str.strptime(pl.Datetime, '%Y-%m-%d %H:%M:%S'),
    pl.exclude('DateTime')
)

print(df)
df.write_parquet("test.parquet", compression='zstd')

Note that compression='zstd' is the default.

Generated Parquet file: test.parquet.zip.

Parquet.jl code

import Pkg
Pkg.activate(temp=true)
Pkg.add([
    Pkg.PackageSpec(name="Parquet", version="0.8.4"),
    Pkg.PackageSpec(name="DataFrames", version="1.5.0")
])
Pkg.status()

import Parquet
Parquet.read_parquet("test.parquet") |> DataFrame

Parquet.jl error

ERROR: zstd error
Stacktrace:
  [1] changemode!(stream::TranscodingStreams.TranscodingStream{CodecZstd.ZstdDecompressor, IOBuffer}, newmode::Symbol)
    @ TranscodingStreams ~/.julia/packages/TranscodingStreams/MQucZ/src/stream.jl:742
  [2] callprocess(stream::TranscodingStreams.TranscodingStream{CodecZstd.ZstdDecompressor, IOBuffer}, inbuf::TranscodingStreams.Buffer, outbuf::TranscodingStreams.Buffer)
    @ TranscodingStreams ~/.julia/packages/TranscodingStreams/MQucZ/src/stream.jl:668
  [3] fillbuffer(stream::TranscodingStreams.TranscodingStream{CodecZstd.ZstdDecompressor, IOBuffer}; eager::Bool)
    @ TranscodingStreams ~/.julia/packages/TranscodingStreams/MQucZ/src/stream.jl:596
  [4] fillbuffer
    @ ~/.julia/packages/TranscodingStreams/MQucZ/src/stream.jl:582 [inlined]
  [5] eof
    @ ~/.julia/packages/TranscodingStreams/MQucZ/src/stream.jl:201 [inlined]
  [6] readbytes!(stream::TranscodingStreams.TranscodingStream{CodecZstd.ZstdDecompressor, IOBuffer}, b::Vector{UInt8}, nb::Int64)
    @ TranscodingStreams ~/.julia/packages/TranscodingStreams/MQucZ/src/stream.jl:387
  [7] readbytes!
    @ ~/.julia/packages/TranscodingStreams/MQucZ/src/stream.jl:384 [inlined]
  [8] (::Parquet.var"#46#47"{Parquet.ColumnChunkPages, Int64})()
    @ Parquet ~/.julia/packages/Parquet/6tj1X/src/reader.jl:187
  [9] (::Parquet.var"#40#42"{Parquet.var"#46#47"{Parquet.ColumnChunkPages, Int64}, Parquet.PageLRU, Tuple{Parquet.PAR2.ColumnChunk, Int64}})()
    @ Parquet ~/.julia/packages/Parquet/6tj1X/src/reader.jl:34
 [10] lock(f::Parquet.var"#40#42"{Parquet.var"#46#47"{Parquet.ColumnChunkPages, Int64}, Parquet.PageLRU, Tuple{Parquet.PAR2.ColumnChunk, Int64}}, l::ReentrantLock)
    @ Base ./lock.jl:229
 [11] cacheget
    @ ~/.julia/packages/Parquet/6tj1X/src/reader.jl:30 [inlined]
 [12] iterate(ccp::Parquet.ColumnChunkPages, startpos::Int64)
    @ Parquet ~/.julia/packages/Parquet/6tj1X/src/reader.jl:167
 [13] iterate(ccpv::Parquet.ColumnChunkPageValues{Int64}, startpos::Int64)
    @ Parquet ~/.julia/packages/Parquet/6tj1X/src/reader.jl:262
 [14] iterate
    @ ~/.julia/packages/Parquet/6tj1X/src/reader.jl:240 [inlined]
 [15] setrow(cursor::Parquet.ColCursor{Int64}, row::Int64)
    @ Parquet ~/.julia/packages/Parquet/6tj1X/src/cursor.jl:114
 [16] Parquet.ColCursor(par::Parquet.File, colname::Vector{String}; rows::UnitRange{Int64}, row::Int64)
    @ Parquet ~/.julia/packages/Parquet/6tj1X/src/cursor.jl:62
 [17] ColCursor
    @ ~/.julia/packages/Parquet/6tj1X/src/cursor.jl:56 [inlined]
 [18] #59
    @ ./none:0 [inlined]
 [19] iterate
    @ ./generator.jl:47 [inlined]
 [20] collect(itr::Base.Generator{Vector{Vector{String}}, Parquet.var"#59#61"{UnitRange{Int64}, Parquet.File}})
    @ Base ./array.jl:782
 [21] Parquet.BatchedColumnsCursor(par::Parquet.File; rows::UnitRange{Int64}, batchsize::Int64, reusebuffer::Bool, use_threads::Bool)
    @ Parquet ~/.julia/packages/Parquet/6tj1X/src/cursor.jl:254
 [22] cursor(table::Parquet.Table)
    @ Parquet ~/.julia/packages/Parquet/6tj1X/src/simple_reader.jl:130
 [23] load(table::Parquet.Table)
    @ Parquet ~/.julia/packages/Parquet/6tj1X/src/simple_reader.jl:134
 [24] getcolumn
    @ ~/.julia/packages/Parquet/6tj1X/src/simple_reader.jl:186 [inlined]
 [25] getcolumn
    @ ~/.julia/packages/Parquet/6tj1X/src/simple_reader.jl:184 [inlined]
 [26] fromcolumns(x::Parquet.Table, names::Vector{Symbol}; copycols::Bool)
    @ DataFrames ~/.julia/packages/DataFrames/LteEl/src/other/tables.jl:36
 [27] fromcolumns
    @ ~/.julia/packages/DataFrames/LteEl/src/other/tables.jl:36 [inlined]
 [28] #fromcolumns#879
    @ ~/.julia/packages/DataFrames/LteEl/src/other/tables.jl:45 [inlined]
 [29] DataFrame(x::Parquet.Table; copycols::Nothing)
    @ DataFrames ~/.julia/packages/DataFrames/LteEl/src/other/tables.jl:59
 [30] DataFrame
    @ ~/.julia/packages/DataFrames/LteEl/src/other/tables.jl:48 [inlined]
 [31] |>(x::Parquet.Table, f::Type{DataFrame})
    @ Base ./operators.jl:907
 [32] top-level scope
    @ REPL[11]:1

Versions

• Parquet v0.8.4
• DataFrames v1.5.0

julia> versioninfo()
Julia Version 1.9.0-rc2
Commit 72aec423c2a (2023-04-01 10:41 UTC)
Platform Info:
  OS: macOS (x86_64-apple-darwin21.4.0)
  CPU: 4 × Intel(R) Core(TM) i5-3330S CPU @ 2.70GHz
  WORD_SIZE: 64
  LIBM: libopenlibm
  LLVM: libLLVM-14.0.6 (ORCJIT, ivybridge)
  Threads: 1 on 4 virtual cores

[ForceBru]

In fact, Parquet.jl can't read any Parquet file produced by Polars 0.17.2 and earlier versions.

Polars code (mostly same as above)

import polars as pl

df = pl.DataFrame({
    'DateTime': [
        '2018-02-01 00:00:00', '2018-02-02 00:00:00',
        '2018-02-03 00:00:00', '2018-02-04 00:00:00',
        '2018-02-05 00:00:00'
    ],
    'String': ['a', 'b', 'c', 'd', 'e'],
    'Float': [1.2, 3.4, 5.6, 7.8, 9.0],
}).select(
    pl.col('DateTime').str.strptime(pl.Datetime, '%Y-%m-%d %H:%M:%S'),
    pl.exclude('DateTime')
)

print(df)
for algo in 'lz4 uncompressed snappy gzip brotli zstd'.split():
    df.write_parquet(f"test.{algo}.parquet", compression=algo)
    print(algo)

This uses all available compression algorithms (except lzo, which apparently isn't implemented).

Julia code

Parquet.jl can't read any of the produced files, even the uncompressed one.

• lz4 compression:

julia> read_parquet("./test.lz4.parquet") |> DataFrame
ERROR: Unknown compression codec for column chunk: 7
Stacktrace:
  [1] error(s::String)
    @ Base ./error.jl:35
  [2] (::Parquet.var"#46#47"{Parquet.ColumnChunkPages, Int64})()
    @ Parquet ~/.julia/packages/Parquet/6tj1X/src/reader.jl:189
  [3] (::Parquet.var"#40#42"{Parquet.var"#46#47"{Parquet.ColumnChunkPages, Int64}, Parquet.PageLRU, Tuple{Parquet.PAR2.ColumnChunk, Int64}})()
    @ Parquet ~/.julia/packages/Parquet/6tj1X/src/reader.jl:34
  [4] lock(f::Parquet.var"#40#42"{Parquet.var"#46#47"{Parquet.ColumnChunkPages, Int64}, Parquet.PageLRU, Tuple{Parquet.PAR2.ColumnChunk, Int64}}, l::ReentrantLock)
    @ Base ./lock.jl:229
  [5] cacheget
    @ ~/.julia/packages/Parquet/6tj1X/src/reader.jl:30 [inlined]
  [6] iterate(ccp::Parquet.ColumnChunkPages, startpos::Int64)
    @ Parquet ~/.julia/packages/Parquet/6tj1X/src/reader.jl:167
  [7] iterate(ccpv::Parquet.ColumnChunkPageValues{Int64}, startpos::Int64)
    @ Parquet ~/.julia/packages/Parquet/6tj1X/src/reader.jl:262
  [8] iterate
    @ ~/.julia/packages/Parquet/6tj1X/src/reader.jl:240 [inlined]
  [9] setrow(cursor::Parquet.ColCursor{Int64}, row::Int64)
    @ Parquet ~/.julia/packages/Parquet/6tj1X/src/cursor.jl:114
 [10] Parquet.ColCursor(par::Parquet.File, colname::Vector{String}; rows::UnitRange{Int64}, row::Int64)
    @ Parquet ~/.julia/packages/Parquet/6tj1X/src/cursor.jl:62
 [11] ColCursor
    @ ~/.julia/packages/Parquet/6tj1X/src/cursor.jl:56 [inlined]
 [12] #59
    @ ./none:0 [inlined]
 [13] iterate
    @ ./generator.jl:47 [inlined]
 [14] collect(itr::Base.Generator{Vector{Vector{String}}, Parquet.var"#59#61"{UnitRange{Int64}, Parquet.File}})
    @ Base ./array.jl:782
 [15] BatchedColumnsCursor(par::Parquet.File; rows::UnitRange{Int64}, batchsize::Int64, reusebuffer::Bool, use_threads::Bool)
    @ Parquet ~/.julia/packages/Parquet/6tj1X/src/cursor.jl:254
 [16] cursor(table::Parquet.Table)
    @ Parquet ~/.julia/packages/Parquet/6tj1X/src/simple_reader.jl:130
 [17] load(table::Parquet.Table)
    @ Parquet ~/.julia/packages/Parquet/6tj1X/src/simple_reader.jl:134
 [18] getcolumn
    @ ~/.julia/packages/Parquet/6tj1X/src/simple_reader.jl:186 [inlined]
 [19] getcolumn
    @ ~/.julia/packages/Parquet/6tj1X/src/simple_reader.jl:184 [inlined]
 [20] fromcolumns(x::Parquet.Table, names::Vector{Symbol}; copycols::Bool)
    @ DataFrames ~/.julia/packages/DataFrames/LteEl/src/other/tables.jl:36
 [21] fromcolumns
    @ ~/.julia/packages/DataFrames/LteEl/src/other/tables.jl:36 [inlined]
 [22] #fromcolumns#879
    @ ~/.julia/packages/DataFrames/LteEl/src/other/tables.jl:45 [inlined]
 [23] DataFrame(x::Parquet.Table; copycols::Nothing)
    @ DataFrames ~/.julia/packages/DataFrames/LteEl/src/other/tables.jl:59
 [24] DataFrame
    @ ~/.julia/packages/DataFrames/LteEl/src/other/tables.jl:48 [inlined]
 [25] |>(x::Parquet.Table, f::Type{DataFrame})
    @ Base ./operators.jl:907
 [26] top-level scope
    @ REPL[2]:1

• uncompressed

julia> read_parquet("./test.uncompressed.parquet") |> DataFrame
ERROR: BoundsError: attempt to access 42-element Vector{UInt8} at index [4538]
Stacktrace:
  [1] getindex
    @ ./essentials.jl:13 [inlined]
  [2] _read_fixed
    @ ~/.julia/packages/Parquet/6tj1X/src/codec.jl:105 [inlined]
  [3] read_fixed
    @ ~/.julia/packages/Parquet/6tj1X/src/codec.jl:97 [inlined]
  [4] read_plain_values(inp::Parquet.InputState, out::Parquet.OutputState{Int64}, count::Int32)
    @ Parquet ~/.julia/packages/Parquet/6tj1X/src/codec.jl:216
  [5] iterate(ccpv::Parquet.ColumnChunkPageValues{Int64}, startpos::Int64)
    @ Parquet ~/.julia/packages/Parquet/6tj1X/src/reader.jl:284
  [6] iterate
    @ ~/.julia/packages/Parquet/6tj1X/src/reader.jl:240 [inlined]
  [7] setrow(cursor::Parquet.ColCursor{Int64}, row::Int64)
    @ Parquet ~/.julia/packages/Parquet/6tj1X/src/cursor.jl:114
  [8] Parquet.ColCursor(par::Parquet.File, colname::Vector{String}; rows::UnitRange{Int64}, row::Int64)
    @ Parquet ~/.julia/packages/Parquet/6tj1X/src/cursor.jl:62
  [9] ColCursor
    @ ~/.julia/packages/Parquet/6tj1X/src/cursor.jl:56 [inlined]
 [10] #59
    @ ./none:0 [inlined]
 [11] iterate
    @ ./generator.jl:47 [inlined]
 [12] collect(itr::Base.Generator{Vector{Vector{String}}, Parquet.var"#59#61"{UnitRange{Int64}, Parquet.File}})
    @ Base ./array.jl:782
 [13] BatchedColumnsCursor(par::Parquet.File; rows::UnitRange{Int64}, batchsize::Int64, reusebuffer::Bool, use_threads::Bool)
    @ Parquet ~/.julia/packages/Parquet/6tj1X/src/cursor.jl:254
 [14] cursor(table::Parquet.Table)
    @ Parquet ~/.julia/packages/Parquet/6tj1X/src/simple_reader.jl:130
 [15] load(table::Parquet.Table)
    @ Parquet ~/.julia/packages/Parquet/6tj1X/src/simple_reader.jl:134
 [16] getcolumn
    @ ~/.julia/packages/Parquet/6tj1X/src/simple_reader.jl:186 [inlined]
 [17] getcolumn
    @ ~/.julia/packages/Parquet/6tj1X/src/simple_reader.jl:184 [inlined]
 [18] fromcolumns(x::Parquet.Table, names::Vector{Symbol}; copycols::Bool)
    @ DataFrames ~/.julia/packages/DataFrames/LteEl/src/other/tables.jl:36
 [19] fromcolumns
    @ ~/.julia/packages/DataFrames/LteEl/src/other/tables.jl:36 [inlined]
 [20] #fromcolumns#879
    @ ~/.julia/packages/DataFrames/LteEl/src/other/tables.jl:45 [inlined]
 [21] DataFrame(x::Parquet.Table; copycols::Nothing)
    @ DataFrames ~/.julia/packages/DataFrames/LteEl/src/other/tables.jl:59
 [22] DataFrame
    @ ~/.julia/packages/DataFrames/LteEl/src/other/tables.jl:48 [inlined]
 [23] |>(x::Parquet.Table, f::Type{DataFrame})
    @ Base ./operators.jl:907
 [24] top-level scope
    @ REPL[3]:1

• snappy compression, which should be "compatible with older parquet readers".

julia> read_parquet("./test.snappy.parquet") |> DataFrame
ERROR: BoundsError: attempt to access 42-element Vector{UInt8} at index [43]
Stacktrace:
  [1] getindex
    @ ./essentials.jl:13 [inlined]
  [2] _read_fixed
    @ ~/.julia/packages/Parquet/6tj1X/src/codec.jl:105 [inlined]
  [3] read_rle_run(inp::Parquet.InputState, out::Parquet.OutputState{Int32}, count::Int64, bits::UInt8, byt::Int64, read_type::Type{UInt32})
    @ Parquet ~/.julia/packages/Parquet/6tj1X/src/codec.jl:321
  [4] read_hybrid(inp::Parquet.InputState, out::Parquet.OutputState{Int32}, count::Int32, bits::UInt8, byt::Int64; read_len::Bool)
    @ Parquet ~/.julia/packages/Parquet/6tj1X/src/codec.jl:311
  [5] read_hybrid
    @ ~/.julia/packages/Parquet/6tj1X/src/codec.jl:283 [inlined]
  [6] read_levels(inp::Parquet.InputState, out::Parquet.OutputState{Int32}, max_val::Int64, enc::Int32, num_values::Int32)
    @ Parquet ~/.julia/packages/Parquet/6tj1X/src/reader.jl:314
  [7] read_levels_and_nmissing(inp::Parquet.InputState, defn_out::Parquet.OutputState{Int32}, repn_out::Parquet.OutputState{Int32}, defn_enc::Int32, repn_enc::Int32, max_defn::Int64, max_repn::Int64, num_values::Int32)
    @ Parquet ~/.julia/packages/Parquet/6tj1X/src/reader.jl:333
  [8] iterate(ccpv::Parquet.ColumnChunkPageValues{Int64}, startpos::Int64)
    @ Parquet ~/.julia/packages/Parquet/6tj1X/src/reader.jl:275
  [9] iterate
    @ ~/.julia/packages/Parquet/6tj1X/src/reader.jl:240 [inlined]
 [10] setrow(cursor::Parquet.ColCursor{Int64}, row::Int64)
    @ Parquet ~/.julia/packages/Parquet/6tj1X/src/cursor.jl:114
 [11] Parquet.ColCursor(par::Parquet.File, colname::Vector{String}; rows::UnitRange{Int64}, row::Int64)
    @ Parquet ~/.julia/packages/Parquet/6tj1X/src/cursor.jl:62
 [12] ColCursor
    @ ~/.julia/packages/Parquet/6tj1X/src/cursor.jl:56 [inlined]
 [13] #59
    @ ./none:0 [inlined]
 [14] iterate
    @ ./generator.jl:47 [inlined]
 [15] collect(itr::Base.Generator{Vector{Vector{String}}, Parquet.var"#59#61"{UnitRange{Int64}, Parquet.File}})
    @ Base ./array.jl:782
 [16] BatchedColumnsCursor(par::Parquet.File; rows::UnitRange{Int64}, batchsize::Int64, reusebuffer::Bool, use_threads::Bool)
    @ Parquet ~/.julia/packages/Parquet/6tj1X/src/cursor.jl:254
 [17] cursor(table::Parquet.Table)
    @ Parquet ~/.julia/packages/Parquet/6tj1X/src/simple_reader.jl:130
 [18] load(table::Parquet.Table)
    @ Parquet ~/.julia/packages/Parquet/6tj1X/src/simple_reader.jl:134
 [19] getcolumn
    @ ~/.julia/packages/Parquet/6tj1X/src/simple_reader.jl:186 [inlined]
 [20] getcolumn
    @ ~/.julia/packages/Parquet/6tj1X/src/simple_reader.jl:184 [inlined]
 [21] fromcolumns(x::Parquet.Table, names::Vector{Symbol}; copycols::Bool)
    @ DataFrames ~/.julia/packages/DataFrames/LteEl/src/other/tables.jl:36
 [22] fromcolumns
    @ ~/.julia/packages/DataFrames/LteEl/src/other/tables.jl:36 [inlined]
 [23] #fromcolumns#879
    @ ~/.julia/packages/DataFrames/LteEl/src/other/tables.jl:45 [inlined]
 [24] DataFrame(x::Parquet.Table; copycols::Nothing)
    @ DataFrames ~/.julia/packages/DataFrames/LteEl/src/other/tables.jl:59
 [25] DataFrame
    @ ~/.julia/packages/DataFrames/LteEl/src/other/tables.jl:48 [inlined]
 [26] |>(x::Parquet.Table, f::Type{DataFrame})
    @ Base ./operators.jl:907
 [27] top-level scope
    @ REPL[4]:1

• gzip compression

julia> read_parquet("./test.gzip.parquet") |> DataFrame
ERROR: zlib error: incorrect header check (code: -3)
Stacktrace:
  [1] changemode!(stream::TranscodingStreams.TranscodingStream{CodecZlib.GzipDecompressor, IOBuffer}, newmode::Symbol)
    @ TranscodingStreams ~/.julia/packages/TranscodingStreams/5yQuA/src/stream.jl:742
  [2] callprocess(stream::TranscodingStreams.TranscodingStream{CodecZlib.GzipDecompressor, IOBuffer}, inbuf::TranscodingStreams.Buffer, outbuf::TranscodingStreams.Buffer)
    @ TranscodingStreams ~/.julia/packages/TranscodingStreams/5yQuA/src/stream.jl:668
  [3] fillbuffer(stream::TranscodingStreams.TranscodingStream{CodecZlib.GzipDecompressor, IOBuffer}; eager::Bool)
    @ TranscodingStreams ~/.julia/packages/TranscodingStreams/5yQuA/src/stream.jl:596
  [4] fillbuffer
    @ ~/.julia/packages/TranscodingStreams/5yQuA/src/stream.jl:582 [inlined]
  [5] eof
    @ ~/.julia/packages/TranscodingStreams/5yQuA/src/stream.jl:201 [inlined]
  [6] readbytes!(stream::TranscodingStreams.TranscodingStream{CodecZlib.GzipDecompressor, IOBuffer}, b::Vector{UInt8}, nb::Int64)
    @ TranscodingStreams ~/.julia/packages/TranscodingStreams/5yQuA/src/stream.jl:387
  [7] readbytes!
    @ ~/.julia/packages/TranscodingStreams/5yQuA/src/stream.jl:384 [inlined]
  [8] (::Parquet.var"#46#47"{Parquet.ColumnChunkPages, Int64})()
    @ Parquet ~/.julia/packages/Parquet/6tj1X/src/reader.jl:185
  [9] (::Parquet.var"#40#42"{Parquet.var"#46#47"{Parquet.ColumnChunkPages, Int64}, Parquet.PageLRU, Tuple{Parquet.PAR2.ColumnChunk, Int64}})()
    @ Parquet ~/.julia/packages/Parquet/6tj1X/src/reader.jl:34
 [10] lock(f::Parquet.var"#40#42"{Parquet.var"#46#47"{Parquet.ColumnChunkPages, Int64}, Parquet.PageLRU, Tuple{Parquet.PAR2.ColumnChunk, Int64}}, l::ReentrantLock)
    @ Base ./lock.jl:229
 [11] cacheget
    @ ~/.julia/packages/Parquet/6tj1X/src/reader.jl:30 [inlined]
 [12] iterate(ccp::Parquet.ColumnChunkPages, startpos::Int64)
    @ Parquet ~/.julia/packages/Parquet/6tj1X/src/reader.jl:167
 [13] iterate(ccpv::Parquet.ColumnChunkPageValues{Int64}, startpos::Int64)
    @ Parquet ~/.julia/packages/Parquet/6tj1X/src/reader.jl:262
 [14] iterate
    @ ~/.julia/packages/Parquet/6tj1X/src/reader.jl:240 [inlined]
 [15] setrow(cursor::Parquet.ColCursor{Int64}, row::Int64)
    @ Parquet ~/.julia/packages/Parquet/6tj1X/src/cursor.jl:114
 [16] Parquet.ColCursor(par::Parquet.File, colname::Vector{String}; rows::UnitRange{Int64}, row::Int64)
    @ Parquet ~/.julia/packages/Parquet/6tj1X/src/cursor.jl:62
 [17] ColCursor
    @ ~/.julia/packages/Parquet/6tj1X/src/cursor.jl:56 [inlined]
 [18] #59
    @ ./none:0 [inlined]
 [19] iterate
    @ ./generator.jl:47 [inlined]
 [20] collect(itr::Base.Generator{Vector{Vector{String}}, Parquet.var"#59#61"{UnitRange{Int64}, Parquet.File}})
    @ Base ./array.jl:782
 [21] BatchedColumnsCursor(par::Parquet.File; rows::UnitRange{Int64}, batchsize::Int64, reusebuffer::Bool, use_threads::Bool)
    @ Parquet ~/.julia/packages/Parquet/6tj1X/src/cursor.jl:254
 [22] cursor(table::Parquet.Table)
    @ Parquet ~/.julia/packages/Parquet/6tj1X/src/simple_reader.jl:130
 [23] load(table::Parquet.Table)
    @ Parquet ~/.julia/packages/Parquet/6tj1X/src/simple_reader.jl:134
 [24] getcolumn
    @ ~/.julia/packages/Parquet/6tj1X/src/simple_reader.jl:186 [inlined]
 [25] getcolumn
    @ ~/.julia/packages/Parquet/6tj1X/src/simple_reader.jl:184 [inlined]
 [26] fromcolumns(x::Parquet.Table, names::Vector{Symbol}; copycols::Bool)
    @ DataFrames ~/.julia/packages/DataFrames/LteEl/src/other/tables.jl:36
 [27] fromcolumns
    @ ~/.julia/packages/DataFrames/LteEl/src/other/tables.jl:36 [inlined]
 [28] #fromcolumns#879
    @ ~/.julia/packages/DataFrames/LteEl/src/other/tables.jl:45 [inlined]
 [29] DataFrame(x::Parquet.Table; copycols::Nothing)
    @ DataFrames ~/.julia/packages/DataFrames/LteEl/src/other/tables.jl:59
 [30] DataFrame
    @ ~/.julia/packages/DataFrames/LteEl/src/other/tables.jl:48 [inlined]
 [31] |>(x::Parquet.Table, f::Type{DataFrame})
    @ Base ./operators.jl:907
 [32] top-level scope
    @ REPL[5]:1

• brotli compression

julia> read_parquet("./test.brotli.parquet") |> DataFrame
ERROR: Unknown compression codec for column chunk: 4
Stacktrace:
  [1] error(s::String)
    @ Base ./error.jl:35
  [2] (::Parquet.var"#46#47"{Parquet.ColumnChunkPages, Int64})()
    @ Parquet ~/.julia/packages/Parquet/6tj1X/src/reader.jl:189
  [3] (::Parquet.var"#40#42"{Parquet.var"#46#47"{Parquet.ColumnChunkPages, Int64}, Parquet.PageLRU, Tuple{Parquet.PAR2.ColumnChunk, Int64}})()
    @ Parquet ~/.julia/packages/Parquet/6tj1X/src/reader.jl:34
  [4] lock(f::Parquet.var"#40#42"{Parquet.var"#46#47"{Parquet.ColumnChunkPages, Int64}, Parquet.PageLRU, Tuple{Parquet.PAR2.ColumnChunk, Int64}}, l::ReentrantLock)
    @ Base ./lock.jl:229
...

• zstd compression: see initial comment above.