funkjedi

Here's the low-down. -- case 1 Q1 - 0.6s with select * Q1 - 0.5s with columns specified -- repeat for every row: Q2 - 55.1s Q2 - 55.9s -- case 2 Q1 - 0.6s with select * Q1 - 0.5s with columns specified Q2 - 52.1s -- case 3 Q1 - 52.9s (also worth noting this query generated hundreds of 1292 'Truncated incorrect DOUBLE value' warnings)

Here's the low-down. -- case 1 Q1 - 0.6s with select * Q1 - 0.5s with columns specified -- repeat for every row: Q2 - 55.1s Q2 - 55.9s -- case 2 Q1 - 0.6s with select * Q1 - 0.5s with columns specified Q2 - 52.1s -- case 3 Q1 - 52.9s

The change in that commit to rely on information schema has made adminer unusable. The performance of querying the KEY_COLUMN_USAGE table in many cases is just brutal (i.e. mins not ms). I realize this comment is not particularly helpful. It's more so just to have this out there.