Of course, allows "duplicate" rows to be inserted if one (or more) of the fields are NULL:
sqlite> create table x(x int, y int, unique (x,y)); sqlite> insert into x values (1,1); sqlite> insert into x values (NULL,1); sqlite> insert into x values (1,NULL); sqlite> insert into x values (1,1); Error: UNIQUE constraint failed: x.x, x.y sqlite> insert into x values (NULL,1); sqlite> insert into x values (1,NULL); If NULLs need to be considered equal, then you have to use a trigger. If there is no intention to use NULLs then the table definition should say that. > On 21 Oct 2016, at 10:53pm, Rick Kohrs <[email protected]> wrote: > > > All fields would match in an existing record compared to that of a > proposed new record. That help? > > So if two rows had all fields identical except for segment, they are not > identical ? Okay, in that case > > CREATE UNIQUE INDEX him_all > ON himawari_db (dateTime, filename, satID, year, month, day, hour, > minute, band , coverage, region, segment, server); > > This will mean it's impossible to insert the second row. The command > > INSERT OR IGNORE ... > > will then not insert the second row, without causing an error. > > Simon. > _______________________________________________ > sqlite-users mailing list > [email protected] > http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/sqlite-users _______________________________________________ sqlite-users mailing list [email protected] http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/sqlite-users
