Eric Brison <eric.brison@xxxxxxxxxxx> writes: > I use tables with several "int[]" columns. > I use the "@>" operator to filter my data. > To increase speed , i install the "intarray" extension. Now queries are > very fast with the specific indexes (i use "gin__int_ops"). > But, in few columns, i have null values in my intarray and i cannot use the > "@>" operator. Do you really need to use intarray, rather than the built-in gin array indexing features? Yeah, the intarray implementation is probably a shade faster, but it's faster precisely because it doesn't handle cases like null entries. If you're in need of that, my suggestion is to ditch intarray. regards, tom lane
