Skip to content

REF: _cython_agg_blocks follow patterns similar to _apply_blockwise #35632

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 14 commits into from
Aug 12, 2020
Merged
Changes from 8 commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
36 changes: 23 additions & 13 deletions pandas/core/groupby/generic.py
Original file line number Diff line number Diff line change
Expand Up @@ -1027,7 +1027,6 @@ def _cython_agg_blocks(
data = data.get_numeric_data(copy=False)

agg_blocks: List[Block] = []
new_items: List[np.ndarray] = []
deleted_items: List[np.ndarray] = []

no_result = object()
Expand Down Expand Up @@ -1059,8 +1058,7 @@ def cast_result_block(result, block: "Block", how: str) -> "Block":
agg_block: Block = block.make_block(result)
return agg_block

for block in data.blocks:
# Avoid inheriting result from earlier in the loop
def blk_func(block):
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

can you type this

result = no_result
locs = block.mgr_locs.as_array
try:
Expand All @@ -1076,8 +1074,7 @@ def cast_result_block(result, block: "Block", how: str) -> "Block":
# we cannot perform the operation
# in an alternate way, exclude the block
assert how == "ohlc"
deleted_items.append(locs)
continue
raise

# call our grouper again with only this block
obj = self.obj[data.items[locs]]
Expand All @@ -1096,8 +1093,7 @@ def cast_result_block(result, block: "Block", how: str) -> "Block":
except TypeError:
# we may have an exception in trying to aggregate
# continue and exclude the block
deleted_items.append(locs)
continue
raise
else:
result = cast(DataFrame, result)
# unwrap DataFrame to get array
Expand All @@ -1107,21 +1103,35 @@ def cast_result_block(result, block: "Block", how: str) -> "Block":
# is a lie. To keep the code-path for the typical non-split case
# clean, we choose to clean up this mess later on.
assert len(locs) == result.shape[1]
new_blocks = []
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

can you type this

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

can you type this

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

and move this to a higher scope (as used in both parts)

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

We only append to it in this scope, everywhere else we set it directly

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

this is used in 3 blackes (sure you don't append and instead directly assign in the other scopes), but it is the same name. I would simply define it once, then append for each (or leave the assign for the others is ok i guess); it just is much easier to grok the fact that we are defining the same variable.

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

done+green

for i, loc in enumerate(locs):
new_items.append(np.array([loc], dtype=locs.dtype))
agg_block = result.iloc[:, [i]]._mgr.blocks[0]
agg_blocks.append(agg_block)
new_blocks.append(agg_block)
else:
result = result._mgr.blocks[0].values
if isinstance(result, np.ndarray) and result.ndim == 1:
result = result.reshape(1, -1)
agg_block = cast_result_block(result, block, how)
new_items.append(locs)
agg_blocks.append(agg_block)
new_blocks = [agg_block]
else:
agg_block = cast_result_block(result, block, how)
new_items.append(locs)
agg_blocks.append(agg_block)
new_blocks = [agg_block]
return new_blocks

skipped = []
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

can you type this

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

done

new_items: List[np.ndarray] = []
for i, block in enumerate(data.blocks):
try:
nbs = blk_func(block)
except (NotImplementedError, TypeError):
# TypeError -> we may have an exception in trying to aggregate
# continue and exclude the block
# NotImplementedError -> "ohlc" with wrong dtype
skipped.append(i)
deleted_items.append(block.mgr_locs.as_array)
else:
agg_blocks.extend(nbs)
new_items.append(block.mgr_locs.as_array)

if not agg_blocks:
raise DataError("No numeric types to aggregate")
Expand Down