This document assumes knowledge of

• what a crontab is
• how to edit a crontab (crontab -e)
• the format of a crontab line (mm hh dd mm weekday command)

The purpose of this document is to explain how to cope with the percentage sign (%) in a crontab entry.

Usually, a % is used to denote a new line in a crontab entry. The first % is special in that it denotes the start of STDIN for the crontab entry's command. A trivial example is:

* * * * * cat - % another minute has passed

This would output the text

another minute has passed

After the first %, all other %s in a crontab entry indicate a new line. So a slightly different trivial example is:

* * * * * cat - % another % minute % has % passed

This would output the text

 another
 minute
 has
 passed

Note how the % has been used to indicate a new line.

The problem is how to use a % in a crontab line to as a % and not as a new line. Many manuals will say escape it with a \. This certainly stops its interpretation as a new line but the shell running the cron job can leave the \ in. For example:

* * * * * echo '\% another \% minute \% has \% passed'

would output the text

\% another \% minute \% has \% passed

Clearly, not what was intended.

A solution is to pass the text through sed. The crontab example now becomes:

* * * * * echo '\% another \% minute \% has \% passed'| sed -e 's|\\||g'

This would output the text

% another % minute % has % passed

which is what was intended.

This technique is very useful when using a MySQL command within a crontab. MySQL command can often have a % in them. Some example are:

• SET @monyy=DATE_FORMAT(NOW(),"%M %Y")
• SELECT * FROM table WHERE name LIKE 'fred%'

So, to have a crontab entry to run the MySQL command

mysql -vv -e "SELECT * FROM table WHERE name LIKE Fred%'" member_list

would have to appear in the crontab as

echo "SELECT * FROM table WHERE name LIKE 'Fred\%'" | sed -e 's|\\||g' | mysql -vv member_list

Pulling the crontab entry apart there is:

1. the echo command sends the MySQL command to STDOUT where it is piped into
2. sed which removes any back slashes before sending the output to STDOUT where it is piped into
3. the mysql command processor which reads its commands from STDIN