I am adding a new table from a Google Sheet to an existing, working app. The Google Sheet has a lot of rows, about 300k, but only 5 columns. Column headers are bold. Nothing unusual, structure is straightforward and consistent.
When I attempt to add the data source, the app becomes non-runnable and I get the error “No columns found Tables must specify a column structure. For some reason, the app definition has been corrupted. It may also be possible that you left your column structure in an inconsistent state. Delete and re-add the table to create the column structure.” …What gives? I’ve tried regenerating the schema, deleting and re-adding the sheet, nothing helps.
Ah, I found a more specific error. Another error says it exceeds the limit of 100k rows. This is weird because Appsheet documentation implies there is no ceiling to the number of rows.
As pointed out by post highlighted by @Shourya_Tomar . You CAN have more rows in your spreadsheets as long as the data size does not hit Google’s limitations.
AppSheet may still stop at loading more than 100,000 rows but if you limit that by using Security Filters, it likely doesn’t become an issue.
Another alternative is to use partitioning if you have a logical way data can be divided between multiple sheets. For example, I have seen some who do so by Year.
Thanks @WillowMobileSys and @Shourya_Tomar . I followed the link to the older article and the “bait and switch” method of adding the data worked like a charm. The sync time isn’t even that bad.
I referred to the limits on data size page (a lot of my apps use large data sets, so I’ve had to look at it, many, many times. We are guilty of using a low-cost, rapid-development web app tool for what should be enterprise-level solutions)… However it’s ambiguous at best. In one place it says there’s a limit, in others it implies, as you said, only Google Drive/Sheets limits apply. It seems fuzzy and poorly defined, and in reality, it seems like the editors attempts to draw a line in the sand, but in reality, it’s more a matter of what an individual application can tolerate depending on the total amount of processing overhead that occurs every sync cycle…. Which is probably tough to quantify and a bit subjective, hence the fuzziness…?
