0 votes
We have a weird issue / bug related to Jupyter Notebooks on Dataiku. We have a dataframe with 74 rows and when you export it through the "Export this dataframe" option you get all the 74 rows as expected. However, one line afterwards we just say df2 = df.reset_index() and then when you try to download df2 (still with 74 rows in the preview) the downloaded file only contains 32 rows and then stops.

Is there something we're missing or is this a known bug?
I think this is related to the string encoding. At row 33 we have a cell that contains an accent. That's where it seems to break. Any ideas on how to avoid this?
How is your text column encoded? Is it explicitly encoded as utf8?
How can one check the encoding of the column of a dataset? We have a string column that contains an accent and the preview of the table in the notebook looks ok, but once you hit "Export this dataframe" the export stops at the row that has the accent in it.

Any suggestion how we can test / modify the workflow / notebook?
Hi, I cannot reproduce this issue on my sample text dataset on Dataiku latest version (4.3.2). Which version of Dataiku are you using? You can check it in the top menu bar > (?) button > About. Would you be willing to attach your original dataset to this thread so I can test it on my computer?

Please log in or register to answer this question.

1,299 questions
1,327 answers
11,867 users

┬ęDataiku 2012-2018 - Privacy Policy