|
Q and A
Asked and Answered
One on one with Bonnie Baker
In
the last "Programmers Only" column
,
Bonnie
Baker posed the question, "What [is] the meaning of the term 'non-sargable'?"
Reader Charley Bontempo ventures some background on the
origins of the word.
Stage 1 or Data Manager query processing was originally performed by the RSS
(Research Storage System), so named by the System R research folks. The DB2
folks changed the name soon after the product release. I believe "sargable"
is a contraction of "search arguable." I never got (or sought) an
explanation of why they decided to contract and use the latter term to label
those predicates that would be handled by the RSS, closest to where the data
is stored. I suppose an analogy is
"timeron," which is a contraction of
"timer on" to specify a unit of query processing/search time, which the
research folks also invented for System R. (They were and still are a great
group!).
Thanks for the first-rate articles; I don't find many of such high quality
around these days.
Charley Bontempo
Yorktown Hgts., NY
Bonnie Baker responds:
Thanks for the historical information, Charley. I had been told that
"sargable" was a IBM-made-up contraction of sorts
for "search-argumentable"
but had not heard the origin for "timeron". I usually tell
folks that the only thing the timeron is good for is picking lottery
numbers. Also, thanks for the kind words about the column.
See a
complete archive of reader/author Q&As
.
Back to
Predicate Evaluation: Part I
by Bonnie Baker
|