Calling column 'name' results in blank table

[wardy912]

Hi,

 This isn't a big issue, but wondering if anyone has seen this before and can shed any light?

I have a notebook that has been running fine, it joins a couple of tables, nothing complicated.

I changed a few column names, one of them to 'name'. For some reason, the table data was removed by a recent run.

In testing, the notebook runs successfully. After deleting the table from the lakehouse, it publishes with the correct columns but no data. All dataframes are correct in the notebook when using display().

After numerous tests, I found that changing the column name to 'sys_name' allowed the data to be successfully published!

Strange issue, interested to hear your thoughts on it.

Thanks

[v-sgandrathi · Accepted solution]

Hi @wardy912,

This issue is likely happening because 'name' is a reserved keyword. Many data processing systems, including Microsoft Fabric, Spark, SQL, and Delta Lake, treat words like 'name', 'type', 'value', or 'id' as special identifiers.

Although your notebook runs and displays data correctly with display(), the problem probably occurs during the publish or write to lakehouse step. The engine may skip or fail to write the data due to schema validation errors caused by using 'name' as a column name.

Renaming the column to something more specific, like 'sys_name', avoids this conflict and allows the table to publish with all data.

To prevent similar issues, try not to use reserved or generic column names such as name, type, id, value, date, or timestamp. Instead, use more descriptive names like user_name, sys_name, or event_date for better compatibility in data processing and publishing.

Thank you and continue using Microsoft Fabric Community Forum.

[wardy912]

Thanks for the advice!