* Frank Bax <fbax@stripped> [040615 12:59]:
> At 03:14 PM 6/15/04, Tim Johnson wrote:
> >We are importing data from 3 CSV files.
<..> >the 3 tables into 1.
> >Taken as a sum, we should have about 330 columns
> >and 5000+ rows.
> Are you saying:
> each CSV contains about 110 columns for the same key values
> - or -
> each CSV contains 1700+ rows for the same 330 columns.
> - or -
> something else?
> In the first case, I would consider loading three tables, then using sql to
> "join" the data by key values and dump it into a new table with all 330
> columns. Then delete the three temp tables.
> In the second case, I would concatenate the CSV files, then import them
> into a table of 330 columns.
> What performance penalty are you expecting with either of these scenarios?
I didn't anticipate any penalty myself (for scenario 1), but was
looking for a second opinion.
I like your tip on using "join". I always like mysql to do the
"heavy lifting" (they're so much better programmers than I)
Tim Johnson <tim@stripped>