On 4 May 2012, at 4:23pm, peter korinis <kori...@earthlink.net> wrote:
> I have scaled down the attributes of interest to 46 columns (discarding the > other 550). No columns are calculated. No updates to this file ... one user > ... only query, sort, etc. type transactions. > So I want to load two 22GB csv files into an empty 46 column table. (I > intend to test load with 999 records by 46 col file.) initially I only have > 1 index on a record # ... am not positive several other fields that I want > to index may not be missing data in some records (I assume that will error > out if I make those an index). > After I get the data loaded and inspect for nulls in prospective index > attributes, can I add indices? Yes. You have done 'CREATE TABLE'. The import process will execute many 'INSERT' commands. And you can 'CREATE INDEX' commands after your data is already present in the table. It will work the way you want. > I was planning to load using sqlite3 CLI ".import" command. Is there a way I > can monitor the progress of the load, with only minimal impact on > performance ? Using File Explorer or the Finder or whatever, just keep an information window on the database file open. You should see the file size increasing constantly while the command is working. Alternatively you might use Task Manager or Activity Monitor to monitor something else that's going on. For example, how many octets that process has written to disk. > I've started several loads only to find out hours later that > nothing has been loaded. Yes, that's very annoying. You keep thinking "It must be finished soon.". Simon. _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
