https://www.postgresql.org/docs/current/tablefunc.html#TABLEFUNC-FUNCTIONS-CROSSTAB-TEXT https://www.postgresql.org/docs/current/tablefunc.html#TABLE... Given

https://www.postgresql.org/docs/current/tablefunc.html#TABLEFUNC-FUNCTIONS-CROSSTAB-TEXT https://www.postgresql.org/docs/current/tablefunc.html#TABLE... Given what this function is doing, a limit on the output seems reasonable. Others have mentioned 1600 columns, which seems like a bit of a strange number, but might correlate internally to a maximum tuple of pointers to the source data. In a quick skim of various searches I don't see any good alternatives to crosstab (AKA pivot) for such monstrously large use cases. Nearly all the toy examples include months, or possibly days where the aggregate inputs collect summary statistics. At 1600+ columns, it's probably better to re-process the data and convert it from tabular to columnar format in whatever way best fits the particular data.