apache / apache/arrow-julia

ACCESS_VIOLATION reading Arrow record batch

Open
#240 0 comments 0 reactions 0 assignees View on GitHub
Dominant language
Julia
Stars
312
Forks
78
PR merge metrics
No merged PRs in 30d

Description

I get this access violation when I read the attached Arrow file. I suspect there is an empty record batch in the Arrow file:
```
Please submit a bug report with steps to reproduce this fault, and any error messages that follow (in their entirety). Thanks.
Exception: EXCEPTION_ACCESS_VIOLATION at 0x5fb26056 -- uncompress at .\boot.jl:0
in expression starting at E:\Source\PIPCore.jl\src\Examples\browsearrow.jl:11
uncompress at .\boot.jl:0
buildbitmap at C:\Users\kobus.herbst\.julia\packages\Arrow\PQ5Mm\src\table.jl:416
build at C:\Users\kobus.herbst\.julia\packages\Arrow\PQ5Mm\src\table.jl:566
unknown function (ip: 000000005fb28cc4)
build at C:\Users\kobus.herbst\.julia\packages\Arrow\PQ5Mm\src\table.jl:402
unknown function (ip: 000000005fb28158)
iterate at C:\Users\kobus.herbst\.julia\packages\Arrow\PQ5Mm\src\table.jl:378
unknown function (ip: 000000005fb21d3d)
copyto! at .\abstractarray.jl:847
_collect at .\array.jl:608 [inlined]
collect at .\array.jl:602 [inlined]
macro expansion at C:\Users\kobus.herbst\.julia\packages\Arrow\PQ5Mm\src\table.jl:288 [inlined]
#82 at .\threadingconstructs.jl:169
unknown function (ip: 000000005fb22363)
jl_apply at /cygdrive/c/buildbot/worker/package_win64/build/src\julia.h:1703 [inlined]
start_task at /cygdrive/c/buildbot/worker/package_win64/build/src\task.c:839
Allocations: 27852364 (Pool: 27841941; Big: 10423); GC: 29
```
This is the code:
```
using Arrow
using DataFrames
using TableBrowse

function browsearrow(path::String, file::String)
df = Arrow.Table(joinpath(path,"$(file).arrow")) |> DataFrame
#TableBrowse.browse(df)
println("Rows $(nrow(df))")
end

browsearrow("D:\\Data\\Demography\\AHRI\\DayExtraction","IndividualResidencyDays_batched")

```
[IndividualResidencyDays_batched.zip](https://github.com/JuliaData/Arrow.jl/files/7193829/IndividualResidencyDays_batched.zip)
The Arrow is in the zip archive

Contributor guide

No contributing guide indexed for this repository

Research direction

Reproduce the crash with Arrow.Table(... ) |> DataFrame using the attached IndividualResidencyDays_batched.arrow file. Start by inspecting src/table.jl around buildbitmap (line 416), build (lines 402 and 566), and iterate (line 378). Done means the file can be read without an access violation and the row count is printed; add coverage for the suspected empty record batch if the cause is confirmed.

Written by the indexing model from the issue text.

Assessment

Tech stack
julia
Domain
data
Issue type
Bug
Difficulty
3/5
Estimated time
1-2 days
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
42/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.