In process of upgrading company computers we had to rewrite a data source for a mail merge being done in MS-Word and MS-Access. Once the data source had been changed from a MS-Access database to MS-Excel all seemed well --accept for the numeric fields. The "smart" default for Excel takes longish numbers and automatically transposes them to scientific notation. Dollar amounts as they appeared in the MS-Word automatically went from 2 digit precision to a floating point number(actually, it looked as if it turned it into an irrational). Any subsequent tweak to the data to get it to _look_ like text appeared doomed. Tried pre-pending an apostrophe, tried suffixing a space, etc. Some of these were successful, but, it meant showing the extraneous character on final output. Granted a pre-pended '#' on long numeric data and '$' on monetary may have been acceptable -- it just should _not_ have to be. Ultimately, layering the numeric data in '=text(' || field || ',"0")' and '=text(' || fieldb || ',"0.00")' worked. This adventure coupled with mail merge seeming to never remember the path to it's data source has led me to believe I killed one two many trees back in the days when I was a COBOL programmer. The IT universe seeks revenge. Or, perhaps it is just wants to slow us down. Every hour debugging is another hour not producing letters.
On my way out the door my co-worker tried to cheer me with the news that the next department to get new computers also require mail merge conversions.