Excel uploading is not working as expected
Strange things happens trying to upload simple excel file (.xlsx) to Dataiku from local computer. I will try to clearly explain how did I try to upload file and what went wrong:
- I click "+Dataset" button and select "Upload your files".
- Drag simple excel file with two columns.
- In "Format/Preview" window I select "skip first lines" to 1 (to skip table names).
- In "Schema" window I select new column name and formats I want. Table looks like as I expected in "Format/Preview" window.
- Without any more changes in "Partitioning" and "Advanced" windows I press "CREATE".
- Dataset that is created does not have columns that I specified (it has default columns col_0 and col_1 instead) and format of one of the columns is changed to boolean (Text is ecpected in both).
I tried to redo the same procedure with .csv file and everything worked fine.
Answers
-
Could you provide a sample of the Excel file?
-
I sent excel to you personal email (not sure if it's possible to upload file here...)
-
Thanks, I tested it using the setting below:
It is indeed a bug for Excel files when renaming the columns at the upload stage in the Settings menu. While we fix it, we recommend using a Prepare visual recipe to perform the renaming. -
I have the same problem.
-
Turribeach Dataiku DSS Core Designer, Neuron, Dataiku DSS Adv Designer, Registered, Neuron 2023 Posts: 1,979 Neuron
This thread is from 2017. Please raise a new thread and explain your issue properly.
-
My problem is exactly the same than this thread from 2017... i read "until we fix it", today nothing has been settled
-
Turribeach Dataiku DSS Core Designer, Neuron, Dataiku DSS Adv Designer, Registered, Neuron 2023 Posts: 1,979 Neuron
Well the thread gives an alternative solution. if you still face the issue on the latest version of Dataiku then clearly it hasn’t been fixed. In which case you can use the alternative solution suggested or if you can’t then you can raise a thread explaining why you can’t use the alternative solution. Either way you should raise a new thread because only the original poster of the thread can mark questions as Solution Accepted. And if you post in the same thread you are preventing questions from being marked as resolved. People are also less likely to respond to an old thread or a thread that has many responses as they need to read the whole thread to understand what is talking about. You can link to this post in your new thread for reference.