[jlanier03]

I'm trying to come up with a shell script that will select a count from a table, compare that to a specific number, then based on the compared results call another script. Kind of like this


select count(*) from tableA
if (count from above) > 500000
then execute scriptsB

any ideas?

[Squashman]

I would need to see the table to make any attempt at this.

[codejockey]

A lot depends on what database/SQL interface you are using. You might consider (1) using a here document to run your SQL query or (2) (if your command-line SQL interface supports it), using back-tics to run the command and collect the output. Some RDBMSs will not accept stdin that is piped from another process, so you can't use a construct such as: cat command-file | sql -- even though the same RDBMS is perfectly happy with the same commands entered interactively (Sybase behaves this way, for example). In such cases, you may be able to use a construct such as sql < command-file to execute your sql commands. If so, you can always write the output to a file (e.g., sql < command-file > sql.out) and parse the output file contents in your script.

Hope this helps. More info appreciated.

[codejockey]

Some RDBMSs will not accept stdin that is piped from another process, so you can't use a construct such as: cat command-file | sql -- even though the same RDBMS is perfectly happy with the same commands entered interactively (Sybase behaves this way, for example).

I sit corrected. As it happens, today I discovered through a bit of experimentation that Sybase will, in fact, accept commands in a pipeline, but it's a bit of work: you must embed the newline character at the proper location(s) in the commands contained in the pipeline. For example, to display information about the columns (fields) of a table named "employee", you might use:

echo "sp_columns employee\ngo" | sql ....

Just wanted to set the record straight.