Currency Reg Expression

[tcerv79]

Currently the regular expression we are using to define currency is not returning results where we expect. If we receive a currency that has text in it, this does not fail. Does anyone have a better reg exp they can suggest to replace the below?

if(Currency !~ /^((|-)[0-9]+\.[0-9][0-9])$/ && Currency !="" && Currency !~ "^[ \t]*$" && Currency !~ "^[ \t]")
          {BadCurrencyCount[x]++;       #Counts number of bad currencies with respect to each column
           BadCurrencyCounter++;	#Counts total number of bad currencies, used for pass/fail check
           BadCurrencyArray[BadCurrencyCounter]=Currency; #Keeps examples of bad currencies
CurrencyColumn[CURRENCYFIELDS[x]] = (BadCurrencyCount[x] "," CURRENCYFIELDS[x]); #Done to link the actual column # to the array subscript

 

[Annihilannic]

While I can see some improvements (there is some redundancy there) it does actually seem to work for me except for the case where the field begins with spaces or tabs; can you give some examples of values that pass the test when they shouldn't? These were my tests:

[COLOR=red] is good[/color]
0.10 is good
[COLOR=red] fred is good[/color]
125.0e is counterfeit!
.03 is counterfeit!
-0.10 is good
fred is counterfeit!
12a.03 is counterfeit!
+0.10 is counterfeit!
[COLOR=red] 123.0 is good[/color]
a125.08 is counterfeit!
0.1 is counterfeit!
100.10 is good
123.0 is counterfeit!

Currency != "" is redundant because Currency !~ "^[ \t]*$" catches that case too. Also (|-) can be replaced by simply -? (which is also more portable), and I would combine all of these into one expression ultimately, but can you clarify the treatment of white space; are you happy if the field is empty? If it contains tabs and/or spaces is it considered empty? And if it is a valid value prefixed by tabs and spaces, is that fine?

Annihilannic.

 

[tcerv79]

Thank you for pointing out the redundancy. That helps! What is the difference then between (|-) and -? , as I am unsure. Essentially if the field NEEDS to be in this format 123.00 or -123.00 anything else(blanks or empty strings or whitespace, $ within the field, or a-zA-Z TEXT) should NOT be in the field.

 

[Annihilannic]

? means that the previous expression 0 or more times. There's no difference between (|-) and -? except that awk on some platforms (e.g. HP-UX, Solaris) doesn't support (|-); I guess they don't like the "empty" expression.

You haven't given an example of a string that passes your regexp test when it shoudln't?

Annihilannic.

 

[feherke]

Hi

? means that the previous expression 0 or more times.

I am sure this was only a typo.

Just to not confuse the beginners :

[tt]r* matches r repeated zero or more times.
r? matches r zero or once.[/tt]

Feherke.

http://rootshell.be/~feherke/

 

[Annihilannic]

Err, yes, that's exactly what I meant.

Annihilannic.

 

[tcerv79]

The strings that should PASS this check are as follows:
0.00
-1.23
1.23
10.20
100.23
0.23
-0.23
WHAT SHOULD NOT PASS THIS EXPRESSION IS
000.00
0001.23
-00012.23
-0000.23
0000.43

Any help would be FANTASTIC

 

[Annihilannic]

Try this:

                if (Currency !~ /^(-?([1-9][0-9]*|0)\.[0-9][0-9])$/ && Currency !~ /^[ \t]*$/ && Currency !~ /^[ \t]/) {

Annihilannic.