I have a dashboard that contains a dataset made up of biweekly CSV uploads. All CSVs contain the same headers and semantic setup. Up until today, when I have uploaded a new CSV, the dashboard would display data from all previous file as well as the new file. Today, after follow the same upload steps as I always have, my dashboard only displays data from today’s CSV upload.
When I click the data source and select “Edit Connection”, I can see the new file along with all other past files; however, the dashboard is only displaying today’s upload data.
I have confirmed that all headers are identical to past CSV files and there have been no datatype changes.
I have nearly the exact same symptoms. For me, at least, It seems to have coincided with the new interface for editing/uploading csv’s. I also noticed it says at the top of the page “2GB out of 2GB free”, which is definitely not true—I’m usually running close to out of space. All my old files are still listed there, just not loading into my reports.
Yes! I, too, was wondering if it might be related to the updated CSV File Upload UI. Thanks for expressing your shared experience. I hope this bug gets sorted out soon!
I also noticed it broke a bunch of my custom formulas because it changed some of my fields from number to text format after re-importing my csv’s into the same dataset.
I too am seeing this and it does seem linked to the new UX to me. I have a dashboard that merges weekly CSVs and now my dash is only showing data from the most recent upload.
You could rather have a google sheet connected as a source, and append it with your new data, if the problem persists. this is what I’ve done and using a service account and python it is basically on Autopilot.
I have the same issue. I’ve noticed that after adding a CSV file, it doesn’t add to the existing bucket, but creates a completely new one. As a result, the report displays data only from the last file. Attempting to add the file directly to the bucket displays an error, but there’s no way to determine the cause. Does anyone know where and how I can report the error to support?
This has not yet been rectified for me, but I found that if I add multiple files NOW it is merging JUST THE NEW ONES… so if I don’t see a fix soon I might need to just reupload all the CSVs needed in this merge.
I am having the same issue. Even when I try re-uploading the same files from the past, Looker Studio seems to no longer recognize the date field as dates…It does seem to be linked to the new UI.
I’ve run into this issue before, and in my case it happened when the dashboard stopped automatically unioning the older CSV files with the newest one. Even though all the files were still listed in the data connection panel, just like you described, the dashboard logic ended up only querying the latest upload.
A few things I’d check based on what fixed it for me:
Verify that all files are actually included in the union/join.
Even if they appear in the connection panel, sometimes one file becomes the “active” source and the rest get dropped from the union without any warning.
Check for hidden formatting mismatches.
I once had a single trailing space in a header name (e.g., Date instead of Date). Everything looked identical, but it caused the dashboard to treat the newest CSV as a separate table.
Check filters on the dashboard.
My dashboard once auto-applied a date filter after importing a new CSV, so it hid all older data even though it was technically still connected.
Refresh extracts / reload the model.
If you’re using an extract or cached dataset, refreshing usually pulls all files back into the view.
In my case it turned out to be a tiny header formatting issue that I didn’t spot until I opened the file in a plain text editor. Hopefully yours is something just as simple.
Same issue here. When I “view on cloud” the obvious difference is, that new files are uploaded as “application/octet-stream” instead of “text/csv“. Also the file suffix is gone, which originally should end with “.csv” in my case.
Seems to be an issue with the new upload interface. Really annoying since it breaks our workflow and we are unable to update our dashboards. To make it worse - the new files should replace earlier versions with a whole year of data that is now gone, because the earlier version was already manually deleted and could not be replaced with the new data now….
Please provide a fix as soon as possible. Thank you!
We have a fix in place and will be rolling it out tomorrow. There will be no action needed on your end and all the files will be queried as opposed to only the last few ones.
Is the fix actually implemented? I am still running into the same issue when I try to upload a new csv file. (e.g. when I “view on cloud” - the existing/old files are in a separate bucket vs. the new file I’m trying to upload. Also the existing/old files are “text/csv” while the new file still has the file type as “application/octet-stream”.) When I upload the new file anyway, it breaks my existing dashboards now instead of just pulling in that single newly uploaded file’s data.