sed & awk, 2nd Edition

phonebill—Track Phone Usage

Contributed by Nick Holloway

The problem is to calculate the cost of phone calls made. In the United Kingdom, charges are made for the number of “units” used during the duration of the call (no free local calls). The length of time a “unit” lasts depends on the charge band (linked to distance) and the charge rate (linked to time of day). You get charged a whole unit as soon as the time period begins.

The input to the program is four fields. The first field is the date (not used). The second field is “band/rate” and is used to look up the length a unit will last. The third field is the length of the call. This can either be “ss,” “mm:ss,” or “hh:mm:ss”. The fourth field is the name of the caller. We keep a stopwatch (old cheap digital), a book, and a pen. Come bill time this is fed through my awk script. This only deals with the cost of the calls, not the standing charge.

The aim of the program was to enable the minimum amount of information to be entered by the callers, and the program could be used to collect together the call costs for each user in one report. It is also written so that if British Telecom changes its costs, these can be done easily in the top of the source (this has been done once already). If more charge bands or rates are added, the table can be simply expanded (wonders of associative arrays). There are no real sanity checks done on the input data. The usage is:

phonebill [ file ... ]

Here is a (short) sample of input and output.

Input ...