kz930 opened a new issue, #8056:
URL: https://github.com/apache/texera/issues/8056

   ### What happened?
   
   The operator puts the two petal values straight into a numpy array and 
standardizes them:
   
   ```python
   length = tuple_[petalLengthCmAttribute]
   width = tuple_[petalWidthCmAttribute]
   features = np.array([[length, width]])
   features = ((features - training_features_means) / training_features_stds)
   ```
   
   When either cell is empty, that value is `None`, the array's dtype becomes 
object, and the subtraction raises `TypeError: unsupported operand type(s) for 
-: 'NoneType' and 'float'`. The run ends, and the message names neither the 
column nor a way out. Nothing in the configuration can exclude the row.
   
   The three neighbouring Hugging Face operators (Sentiment Analysis, Spam 
Detection, Text Summarization) were changed in #7549 to keep the row and leave 
the result columns empty instead. This operator reads numeric columns rather 
than a text column, so that change did not reach it, and it is now the one 
operator in the group that ends the run.
   
   I expected it to answer the empty cell the way its three siblings do: keep 
the row, leave the prediction columns empty.
   
   ### How to reproduce?
   
   1. CSV File Scan on a table whose petal length or petal width column has a 
blank cell.
   2. Connect Hugging Face Iris Logistic Regression, pointing its two 
attributes at those columns.
   3. Run.
   
   The run ends with the TypeError below. With every cell filled it completes 
normally.
   
   ### Version/Branch
   
   1.3.0-incubating-SNAPSHOT (main)
   
   ### Relevant log output
   
   ```shell
   TypeError: unsupported operand type(s) for -: 'NoneType' and 'float'
   ```
   


-- 
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]

Reply via email to