[
https://issues.apache.org/jira/browse/ARROW-16000?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17569009#comment-17569009
]
Joost Hoozemans commented on ARROW-16000:
-----------------------------------------
Thanks everyone for the advice. What makes CsvFragmentScanOptions the preferred
place over csv.ReadOptions? CsvFragmentScanOptions right now doesn't directly
store any properties itself, it only carries a csv.ConvertOptions and
csv.ReadOptions. And compression and encoding sound like properties of a whole
file, not a fragment (although I don't know if that is what Fragment means
here).
Would it make sense as first attempt for me to add a TransformInputStream to
CsvFragmentScanOptions or ReadOptions? Because then I can create 1 in python in
the same way read_csv does it (with MakeTransformInputStream, with a callback
into en/decode functions in a python library). Then we can see if there is a
performance problem. Then later we could add functionality that creates a
TransformInputStream in c++ world with a callback to some external library as
Antoine suggested
> [C++][Dataset] Support Latin-1 encoding
> ---------------------------------------
>
> Key: ARROW-16000
> URL: https://issues.apache.org/jira/browse/ARROW-16000
> Project: Apache Arrow
> Issue Type: Improvement
> Components: C++
> Reporter: Nicola Crane
> Assignee: Joost Hoozemans
> Priority: Major
>
> In ARROW-15992 a user is reporting issues with trying to read in files with
> Latin-1 encoding. I had a look through the docs for the Dataset API and I
> don't think this is currently supported.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
