Chances are, you should have approached your data differently. What are those 50 columns? Is it something like dates? Because in a case like that you should’ve designed your table as one row per cell (with a column for a date) and group by date in top.
Can’t find a more recent implementation of this but here’s the idea:
If this doesn’t work and you actually need to combine the columns, one way to do this is with a custom pack. The pack would read a row from the Coda API and return all cell values as a single array of strings.
Finally, there’s black magic. But it goes far beyond “simple”.
Even if those are attributes, it could make a lot of sense to have a separate table of DB Project attributes with columns: Project, Attribute, and Value. So your row-of-50-columns will be the 50 rows on that table, and you’ll have the flexibility of aggregating over those rows normally with e.g. a Filter