You are viewing a plain text version of this content. The canonical link for it is here.
Posted to jira@arrow.apache.org by "Marc Garcia (Jira)" <ji...@apache.org> on 2022/09/18 17:22:00 UTC
[jira] [Commented] (ARROW-17766) Add option to only load specific columns from csv
[ https://issues.apache.org/jira/browse/ARROW-17766?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17606331#comment-17606331 ]
Marc Garcia commented on ARROW-17766:
-------------------------------------
Ok, my bad, seems like it was there, `include_columns` at the end of the `ConverterOptions` parameters, I couldn't find it after looking at all the options 3 times (just found it checking the source code): https://arrow.apache.org/docs/python/generated/pyarrow.csv.ConvertOptions.html#pyarrow.csv.ConvertOptions
> Add option to only load specific columns from csv
> -------------------------------------------------
>
> Key: ARROW-17766
> URL: https://issues.apache.org/jira/browse/ARROW-17766
> Project: Apache Arrow
> Issue Type: New Feature
> Components: Python
> Reporter: Marc Garcia
> Priority: Major
>
> I may be missing something, but after checking in detail the documentation of pyarrow, I can't find an option equivalent to `pandas.read_csv(..., usecols=['col_to_load_1', 'col_to_load_2'])`.
> This would be useful for example when loading CSV files with lots of columns where only few are needed. Having to load all the information in the CSV seems like a significant waste of resources.
> I guess a parameter `load_columns` could be added to `pyarrow.csv.ReadOptions`.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)