fix: analysis report when there is a column with mixed data types #131
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
johnnygreco
commented
Dec 12, 2025
Comment on lines
-32
to
-37
| @model_validator(mode="after") | ||
| def ensure_pyarrow_backend(self) -> Self: | ||
| if not all(isinstance(dtype, pd.ArrowDtype) for dtype in self.df.dtypes): | ||
| self.df = pa.Table.from_pandas(self.df).to_pandas(types_mapper=pd.ArrowDtype) | ||
| return self | ||
|
|
Contributor
Author
There was a problem hiding this comment.
this validator is what was causing generation jobs to fail at the profiling step
johnnygreco
force-pushed
the
johnny/bug/128-profiling-mixed-dtype-bug
branch
from
b77e803 to
c8cafcf
Compare
johnnygreco
commented
Dec 12, 2025
Comment on lines
108
to
112
| logger.warning( | ||
| "⚠️ Unable to convert the dataset to a PyArrow backend. This is often due to at least " | ||
| "one column having mixed data types. As a result, the reported data types " | ||
| "will be inferred from the type of the first non-null value of each column." | ||
| ) |
Contributor
Author
There was a problem hiding this comment.
here's the clue to the user that their dataset has at least one column with mixed data types
Contributor
There was a problem hiding this comment.
Not blocking this needed bugfix: but is it possible to notify the user about which column generated the issue?
Contributor
Author
There was a problem hiding this comment.
good question, probably can grab it from the exception message. let me check that
Contributor
Author
There was a problem hiding this comment.
new message looks like this
[17:19:19] [WARNING] ⚠️ Unable to convert the dataset to a PyArrow backend
[17:19:19] [WARNING] |-- Conversion Error Message: Conversion failed for column 'nano_response' with type object
[17:19:19] [WARNING] |-- This is often due to at least one column having mixed data types
[17:19:19] [WARNING] |-- Note: Reported data types will be inferred from the first non-null value of each column
eric-tramel
approved these changes
Dec 13, 2025
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
closed #128
The issue here is that we use the pyarrow backend of pandas to infer the data type we report (e.g., it distinguishes different types of objects, whereas pandas labels strings, lists, and dicts as "object" type). However, sometimes we might generate a column with multiple data types (whether it is intentional or not), particularly if the column is generated by an LLM. This makes pyarrow unhappy.
This PR makes it so that if we can't convert the dataframe backend to pyarrow, (1) the dataset generation does not fail and (2) we fallback to using the data type of the first non-null element in the analysis report.