[ https://issues.apache.org/jira/browse/ARROW-9367?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17302404#comment-17302404 ]
Uwe Korn commented on ARROW-9367: --------------------------------- This is now working in 3.0: {code} import pyarrow as pa import pyarrow.compute as pc table = pa.table({ "a": [1, 2, 3, -1], "b": [10, 9, 9, 7], "c": [10, 10, 8, 11] }) indices = pc.sort_indices(table, sort_keys=[("b", "ascending"), ("c", "ascending")]) table = pc.take(table, indices) table.to_pydict() # {'a': [-1, 3, 2, 1], 'b': [7, 9, 9, 10], 'c': [11, 8, 10, 10]} {code} > [Python] Sorting on pyarrow data structures ? > --------------------------------------------- > > Key: ARROW-9367 > URL: https://issues.apache.org/jira/browse/ARROW-9367 > Project: Apache Arrow > Issue Type: Wish > Components: Python > Reporter: Athanassios Hatzis > Priority: Major > Labels: sort > > Hi, I consider sorting a fundamental operation for any in-memory data > structures, including those of PyArrow. > It would be nice if pa.array, pa.table, etc had sorting methods but I did not > find any. One has to pass sorting indices calculated from some other library, > such as numpy, to sort them. Sorting indices could have been calculated > directly from PyArrow. Am I missing something here ? That increases > significantly complexity for the developer. > Do you have any plans on implementing such a feature in the near future ? -- This message was sent by Atlassian Jira (v8.3.4#803005)