[GitHub] [arrow] 0x26res opened a new issue, #15078: add type and mask argument to pa.MapArray.from_array

[GitBox]

0x26res opened a new issue, #15078:
URL: https://github.com/apache/arrow/issues/15078

   ### Describe the enhancement requested
   
   I would like to be able to create MapArray, using the from_array function, 
and:
   1. specify missing value (by that I mean the whole map is null), by passing 
a `mask` argument
   2. specify that each individual map values are not nullable, by passing a 
`type` argument
   
   This would make 
[pa.MapArray.from_array](https://arrow.apache.org/docs/python/generated/pyarrow.MapArray.html#pyarrow.MapArray.from_arrays)
 similar to 
[pa.ListArray.from_array](https://arrow.apache.org/docs/python/generated/pyarrow.ListArray.html#pyarrow.ListArray.from_arrays)
   
   Here's an example of what I am trying to do.
   
   1. Can't create a map array with missing values
   
   Using `pa.array` I can create a MapArray with a null value:
   ```
   array_from_python = pa.array([[], None, [(1, None), (2, "2")]], map_type)
   ```
   
   But I can't do it with `pa.MapArray.from_arrays`, because I can't pass a mask
   ```
   array_from_arrays = pa.MapArray.from_arrays(
       pa.array([0, 0, 0, 2], pa.int32()),
       pa.array([1, 2]),
       pa.array([None, "2"]),
   )
   ```
   
   I could not find a workaround for this one, `replace_with_mask` isn't 
supported for MapArray it seems:
   ```
   pc.replace_with_mask(
       array_from_arrays, array_from_python.is_null(), pa.scalar(None, 
array_from_arrays.type)
   )
   ```
   
   2. Can't pass the type of the map array:
   
   There is a niche use case where one would have to specify the type of the 
array. 
   That's when one wants to specify if individual values of the maps are not 
nullable (or if one want to change the name of the key/value fields of the map):
   
   ```
   array_from_arrays = pa.MapArray.from_arrays(
       pa.array([0, 0, 0, 2], pa.int32()),
       pa.array([1, 2]),
       pa.array(["1", "2"]),
       type=pa.map_(pa.int64(), pa.field("value", pa.string(), nullable=False))
   )
   ```
   
   For the moment, I can workaround this problem by casting the map array.
   
   
   ### Component(s)
   
   Python


-- 
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.

To unsubscribe, e-mail: issues-unsubscr...@arrow.apache.org.apache.org

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org