How do i get the variables to actually get in here....

krujos has asked for the wisdom of the Perl Monks concerning the following question:

Replies are listed 'Best First'.
Re: How do i get the variables to actually get in here.... by wog (Curate) on Nov 07, 2001 at 02:40 UTC
At least part of your problem seems to be with `while (defined(<>))`. The `<>` takes a line from the filehandle, and then you test it for being defined, and (continue to) loop if it is. Unfortunatly, unlike using `while(<>)` it doesn't actually assign the value into `$_`. To do that you can use `while (defined($_=<>))`, but that's exactly equivilent to `while(<>)`. I would advise that you look at using split, instead of regex for this task. It looks like it probably would be better suited for this data.	[reply] [d/l] [select]
Re: Re: How do i get the variables to actually get in here.... by krujos (Curate) on Nov 07, 2001 at 02:53 UTC
Split would be nice and easy but alas my prof has said its "off limits" for this project	[reply]
Re: How do i get the variables to actually get in here.... by dragonchild (Archbishop) on Nov 07, 2001 at 02:34 UTC
.* sucks. Don't use it unless you have to. Better would be to do something like: `my @colNames = qw(name status class major po phone advisor email); while (<>) { chomp; my %hash; @hash(@colNames) = /^(\w+, \w+ (?:\w+)?)\s(\w+)\s .../; print "$hash{name}\n"; print OUTPUT "$hash{name}\n"; } close OUTPUT \|\| die "Cannot close OUTPUT: $!\n";` [download] I didn't complete the regex ... that's left as an exercise for the reader. :-) Update: Fixed `while (defined <>)` as per wog's comment. ------ We are the carpenters and bricklayers of the Information Age. Don't go borrowing trouble. For programmers, this means Worry only about what you need to implement.	[reply] [d/l] [select]
Re: How do i get the variables to actually get in here.... by runrig (Abbot) on Nov 07, 2001 at 02:33 UTC
Your 'name' section is only '\w+' when the name in your data contains more than just '\w' characters. The regex contains '\W\W*' which is equivalent to '\W+', and if you're looking for whitespace between fields you're better off using '\s+' anyway. In fact, you'd do well to use better matching parameters on all of the fields, you're capturing all your fields with '\w+', when, e.g., the phone number contains more than just '\w' characters ('-' for instance), and won't contain any alphabetic '\w' characters. See perldoc perlre for explanations of what '\w' and '\s' are and hints on what you ought to be using instead.	[reply]
variable number of fields by pike (Monk) on Nov 07, 2001 at 13:58 UTC
Actually, part of the problem seems to be your input data. Both 'name' and 'advisor' can consist of more than one name according to your example. Which raises the question: are the names always of the form 'lastname, firstname middlename'? What if a person has no middle name or more than one? In any case I don't see that your regex covers that. Since this is a homework assignment, you probably can't change the input data (which is what I would normally do - use a separator like ':' to delimit the fields), so the way to go is probably: - split at blanks (use `@a = /([\w-,@]+\s+)/`) - check which fields terminate in ',' - these are the last names of the person and his advisor - check how many fields are between the person's last name and the advisor's last name and how many after the advisor's last name to find out how many first / middle names were given in each case Hope this helps you, pike	[reply] [d/l]
Re: How do i get the variables to actually get in here.... by krujos (Curate) on Nov 07, 2001 at 02:30 UTC
there is more than one space between the fields in the input file. sorry about that..	[reply]
Re: How do i get the variables to actually get in here.... by blackmateria (Chaplain) on Nov 07, 2001 at 02:40 UTC
You're using `=` instead of `=~` on the line with the regex. That alone will make the script not work. Look in perlop and perlre for more information on `=~` and regexes, if you're interested. That regex looks awfully complicated to me. I think you should look into using split instead (in perlfunc). I can't say for sure whether this will work without seeing your input data, but it seems pretty likely given that all your regex contains is permutations of \W and \w. One additional point, don't use `while (defined(<>))`. `<>` is always defined until end-of-file, and I don't think that form assigns the result of `<>` to `$_`. Use `while (<>)` instead. Update: turns out `=` will do the right thing in this case, at least according to dragonchild (see next post). Oh well, it was probably the `while (defined(<>))` bit anyway...	[reply] [d/l] [select]
Re: Re: How do i get the variables to actually get in here.... by dragonchild (Archbishop) on Nov 07, 2001 at 02:55 UTC
Actually, he wants to be using =, not =~. =~ says to use the variable on the left as the thing to perform the regex on. If there is no source, `$_ =~` is automatically put in at the left of the regex. As he wants to assign the results of the regex, performed on $_, to the lvalues, = is correct. This is actually a very common thing to do when parsing something that split doesn't work nicely on. Again, that's his example data. So, = and a regex are exactly the right calls here. (I spent 3 hours Friday working with similar data that was almost fixed-delimited, but wasn't, so I had to move away from unpack to a regex and it worked perfectly. Yes, they're complicated, but they work.) (I didn't realize this at the time, but this is my 500th node!) ------ We are the carpenters and bricklayers of the Information Age. Don't go borrowing trouble. For programmers, this means Worry only about what you need to implement.	[reply] [d/l]