forum.vdsworld.com

Skit3000 · Posted: Mon Jul 05, 2004 8:12 am Post subject: Parsing with double quotes

I think I just found a bug (or handy trick) with the parse command. If you put double quotes around the current fieldseperator, it won't be recognized by the parse command:

Dr. Dread · Posted: Mon Jul 05, 2004 8:47 am Post subject:

It would be a bug if it didn't work that way... otherwise you would have problems when using PARSE
for database files where your field sep may occur inside field contents.

Greetz
Dr. Dread

Skit3000 · Posted: Mon Jul 05, 2004 8:51 am Post subject:

I don't really know about that. Because if you have database file which is using single quotes to keep the field contents together, it still doesn't work properly if one of these field contents has a double quote in it. When using database files, I think it's best to make a parsing routine yourself insteed of using the parse command... Smile

Garrett · Posted: Mon Jul 05, 2004 3:39 pm Post subject:

Recently, I started building a little app to help me manage a CSV database
for a small web site that I'm taking over. I decided to go with a CSV that
is as close to the ?known? standards as possible, which means that each
entry is bound by a quote mark, and each entry is seperated by a comma.

jules · Posted: Mon Jul 05, 2004 8:39 pm Post subject:

Hmm. I should have thought the parse command would cope with comma separated data okay. This works as I would expect:

Dr. Dread · Posted: Tue Jul 06, 2004 5:34 am Post subject:

I've been using PARSE for flatfile databases for a long time. Field sep's inside fields haven't
been any problem as long as the fields are quoted (standard double quotes). Double quotes
inside fields may, however, be problematic; I don't allow them in there, replace by single
quotes.

Greetz
Dread

Garrett · Posted: Tue Jul 06, 2004 3:56 pm Post subject:

Unfortunately for me, the data is based on data that I can't change to
fit my needs. If it were so, I would have used @tab() for the seperater
instead. That, and the fact that the csv file is going to be available for
other people to use, so I have to keep it all intact and in proper format.

Honestly though, I don't have a problem using the @pos() loop. In fact,
I kind of feel safer doing it this way than relying on the PARSE command.

Also, if you think about it, using PARSE means that you have to hard code
how many fields will be present, but using @pos() and a list, you can code
a csv program that can be flexible and dynamic.

jules · Posted: Tue Jul 06, 2004 6:23 pm Post subject:

OK. I only brought it up because this has been posted under Bug Reports. The Parse command was really intended to provide a way of getting the information from VDS functions that can return multiple values in one call, like @verinfo, @file etc. It's handy that it can be used for other things, but I don't think that its failure to work with CSV files counts as a bug.

Garrett · Posted: Tue Jul 06, 2004 7:26 pm Post subject:

I agree.

FreezingFire · **Admin Team** Joined: 23 Jun 2002 Posts: 3508

Moved to General Help. Smile