I’m running into a persistent bug in Looker Studio when blending a Google Sheets source with Search Console data. My Google Sheet contains a cleaned list of landing pages (URLs) in a column called Landing Page CLEAN. I’m using this as the join key for blending with Search Console’s Landing Page.
The issue:
Even though the Landing Page CLEAN column clearly contains text (URLs) and works correctly in the spreadsheet (shown as expected via =QUERY({…})), Looker Studio does not recognise it as a dimension. It either:
- Misclassifies the column as numeric (probably due to auto-type inference),
- Or omits it entirely from the dimensions list in the data blend UI.
This makes it impossible to use as a join key, breaking the entire blend and rendering the report non-functional.
What I’ve Tried (none worked so far):
- Wrapping the column in TEXT(…, “@”)
- Using ARRAYFORMULA(TEXT(…))
- Ensuring no empty or numeric-looking values at the top of the column
- Setting explicit column headers and double-checking types in Sheets
- Manually inspecting the column in Looker Studio — still treated as a metric
- Changing the source sheet, range, and row depths
- Trying the blend from scratch — still broken
Querying only a fixed number of rows (to match row lengths)
Adding dummy text URLs in the first row
Notes:
The issue only occurs with blended data, not with visualising the sheet on its own.
I can see and use the column fine in unblended tables.
This appears to be a type inference bug on Looker Studio’s end that misreads text as numbers during import.
Request:
Can anyone help me:
- Force Looker Studio to recognise a sheet column as text, even when it guesses otherwise?
- Work around this bug so I can complete the data blend and visualise correctly?
Open to creative workarounds or any advice from others who’ve dealt with this. Thanks in advance!