ypppxwcxxk-lab opened a new issue, #48182:
URL: https://github.com/apache/arrow/issues/48182
### Describe the bug, including details regarding any error messages,
version, and platform.
RecordBatch.from_struct_array(array_of_lists_of_structs[i].values) returns a
RecordBatch that is somewhat corrupt and seems to contain all structs from the
array (as if the lists had been flattened), not just the rows from the ith
sub-list.
Example code:
```python
import pyarrow
struct_type = pyarrow.struct([("x", pyarrow.int64()), ("y",
pyarrow.int64())])
list0 = [{"x": 1, "y": 2}, {"x": 3, "y": 4}]
list1 = [{"x": 5, "y": 6}]
list2 = [{"x": 7, "y": 8}, {"x": 9, "y": 10}]
list_of_structs = pyarrow.array([list0, list1, list2],
type=pyarrow.list_(struct_type))
# Get the first list - should be [{x:1,y:2}, {x:3,y:4}]
elem0 = list_of_structs[0]
batch_for_elem0 = pyarrow.RecordBatch.from_struct_array(elem0.values)
print("ver", pyarrow.__version__)
print("RIGHT:", len(elem0.values), elem0.values.field("x"))
print("WRONG?:", len(batch_for_elem0), batch_for_elem0["x"])
```
Output:
```
ver 22.0.0
RIGHT: 2 [
1,
3
]
WRONG?: 2 [
1,
3,
5,
7,
9
]
```
### 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: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
