On 11 Jul 2018, at 6:01pm, Randall Smith <rsm...@qti.qualcomm.com> wrote:
> (o) Allow humans to view the contents of a DB without custom tools. > (o) Have a way to see what has changed between V1 and V2 of a database, e.g., > for a "change review." SQL is based around Ted Codd's view of relational databases. One of the fundamentals of this view is that a table of rows has no inherent order. Rows of data in one table are like pebbles in a bag, not dots on a line. Or, if you prefer computing terms, they're a set, not an array. However a text file does have an inherent order. So any tool that coverts a SQL database to a text file must do something arbitrary: pick an order. The conclusion is that if you're comparing two databases you need to compare them while they're databases, not when you're looking at them as text files. So if your first point above comes down to "view the database as text" then the above two points conflict with one-another. You need to do that first point and the second point as two separate procedures, not do step 1 then use the output of that for step 2. As a more direct answer to an earlier question, the ".dump" command of the CLI does dump the data in a predictable and consistent order. But it is not documented to do so. So a future version can change that. Simon. _______________________________________________ sqlite-users mailing list sqlite-users@mailinglists.sqlite.org http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/sqlite-users
