LabVIEW
Regular Expression for match pattern
Hmm, quite right. Now thats annoying behaviour 
Ok. I took a closer look at the different ways to parse out the strings and, while this one may not be the quickest way to get things done, it was able to determine every different kind of delimiter I could think of and got all the data right in my testings. Could you also test this one to see if I missed anything?
Last one I posted had a blatent bug in it. While it did take care of delimiters just fine I forgot to go back and test to make sure plain old comma seperators worked still (it was truncating the first digit of the numbers 
).
).Interesting problem trying to strip off characters for the parsing.
If you strip two at a time delimiters are not an issue but numbers are invalid. If you only strip one at a time numbers are correct but delimiters with minus signs can be misread as negations. I rebuilt it last night to, hopefully, address both of these issues by preparsing the false negatives out. I am still testing with different string inputs but I think this version is solid.
Yes, the user should not be burdened with the task of proper formatting, the program should do its best to extract the data from the dirty string.
(I am always annoyed by e.g. web fields for the credit card # that state "no spaces" or similar. Any code should be able to strip spaces from a CC #!)
Anyway, I prefer "scan strings for tokens". Here you can give a list of all possible separators and consecutive seperators are contracted by default. I this case, you would use separators " " and ",", but nothing prevents you from listing every ASCII value except "-" and 0..9. Expand as needed. A simple example is attached (LAbVIEW 7.1).
Message Edited by altenbach on 09-09-2005 11:14 AM
Here's a simple variation that treats everything except [0..9 or -] as delimiter. It clean up most input strings nicely.
Message Edited by altenbach on 09-09-2005 12:00 PM
