Charged rows per write statement, per table, largest factor first.
The two counters this divides were already being collected, so the factor costs nothing new to
obtain. What it does NOT tell you is which index -- only that a table charges more than it stores,
which is the signal to go and look. A factor of 1.0 on a hot table means there is nothing to win
there, and that is worth knowing before touching a schema.
A table with statements but zero charged rows reports a factor of 0 rather than being dropped: an
all-no-op table is a finding too, because the statements still cost CPU.
Charged rows per write statement, per table, largest factor first.
The two counters this divides were already being collected, so the factor costs nothing new to obtain. What it does NOT tell you is which index -- only that a table charges more than it stores, which is the signal to go and look. A factor of 1.0 on a hot table means there is nothing to win there, and that is worth knowing before touching a schema.
A table with statements but zero charged rows reports a factor of 0 rather than being dropped: an all-no-op table is a finding too, because the statements still cost CPU.