On Wed, 11 Dec 2019, Tom Lane wrote:
String comparisons in non-C collations tend to be a lot slower than they are in C collation. Whether this makes a noticeable difference to you depends on your workload, but certainly we've seen performance gripes that trace to that.
Tom, How interesting.
If your data doesn't require the larger character set of UTF8, then using LATIN-any is going to offer some space savings (for non-ASCII characters) plus minor performance benefits due to the lack of variable-width characters. This is less significant than the collation issue, though, for most people.
I doubt that my use will notice meaningful differences. Since there are only two or three databases in UTF8 and its collation perhaps I'll convert those to LATIN1 and C. Thanks for the insights. Regards, Rich
