westonpace commented on a change in pull request #9323:
URL: https://github.com/apache/arrow/pull/9323#discussion_r565550260
##########
File path: cpp/src/arrow/compute/api_vector.h
##########
@@ -63,6 +63,25 @@ enum class SortOrder {
Descending,
};
+struct DictionaryEncodeOptions : public FunctionOptions {
+ /// Configure how null values will be encoded
+ enum NullEncodingBehavior {
+ /// the null value will be added to the dictionary with a proper index
+ ENCODE,
+ /// the null value will be masked in the indices array
+ MASK,
+ /// the null value will not be included in the dictionary
+ SKIP
Review comment:
Yes, sorry, I will discard the SKIP/IGNORE option. At one point I
thought that was the current behavior and so I added the option. Later testing
showed that I simply misunderstood what was happening and I forgot to fully
remove the option. ENCODE and MASK are the only options.
@bkietz I'm not sure what inline examples of such behavior would mean.
Something like this maybe...
```
ENCODE [5, 5, null] -> Dictionary [5, null] / Indices [0, 0, 1]
MASK [5, 5, null] -> Dictionary [5] / Indices [0, 0, null]
```
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
