trim and REReplace

[cfmx2008]

Hi,

I have this number: 00000000010000000099B I need to remove the 0s before 10 and 99 and the alphabets like B after 99. I also need to replace the 0s between 10 and 99 by "-". I don't know how to use trim and Rereplace for these two numbers in this string. Any Idea? I appreciate it. Thanks

[acoder]

What's the pattern of these strings?

Replacing 0 and B is easy, but the 10 would make it more difficult because it has a 0!

[cfmx2008]

Agree. I have this problem too. The pattern is : 10 digits at the beginning of the string (for the 1st number) + 10 more digits after (for the 2ed number) + 1 alphabet.

I think we should clean the 4 or 5 first digits (0s) from the beginning of the string then leave the next 5 digits then insert our "-" and again clean the 4 or 5 next 0s and leave the rest and again clean the alphabet. What do you think?

[acoder]

That does make the problem easier because you have a consistent pattern.

However, I notice that in the example string you have 9 0s then 10. Is that correct or should it be 8 0s then a 10?

[cfmx2008]

In this example I have 9 0s. but remember the 1st 10 digits are belong to 1st number. In this case number is 1 so we have 9 0s left. The real number could be 111 then we will have 7 0s before the number.

[acoder]

Use something like ^0+([1-9]+)(0+)([1-9]+)([A-Z]?)$ and replace with $1-$3.

I haven't tested this with REReplace, but it should give you an idea. It matches 0s at the beginning followed by a number followed by 0s followed by a number followed by letters.

[cfmx2008]

Cool. I'll try it and let you know.

[cfmx2008]

Here is the code:
<cfset StreetNumber = REReplace(#StreetNumber#,"^0+([1-9]+)(0+)([1-9]+)([A-Z]?)$","$1-$3", "all")>

it will replace the whole StreetNumber by $1-$3. I tried it with trim() as well. It does the same. #StreetNumber# is 00000000010000000099B

[cfmx2008]

I Got the exact pattern:

Street number ranges for near matches are returned in the following fixed-width format:

START RANGE (10 bytes) + END RANGE (10 bytes) + OEB code (1 byte) left-padded with zeros



Secondary (apartment) ranges for near matches are returned in the following fixed-width format:

START RANGE (8 bytes) + END RANGE (8 bytes) + OEB code (1 byte) left-padded with zeros



The OEB is “odd-even-both” code, meaning that the range either contains all Odd numbers, Even numbers, or Both.



In our example,

"00000000010000000099B" means “buildings 1-99 Both odd and even”, i.e. {1,2,3,4,…. 98,99}



As you see, all we need to do is chop off all leading zeros in each of the range blocks.

Please don't give up on me:-)

[acoder]

Sorry, $1-$3 was JavaScript (and possibly other languages) notation for back-referencing. In Coldfusion, it's "\1-\3".

[cfmx2008]

Thanks. But this replacement works only for this example. Did you get the chance to see the exact pattern I've posted? I need the "-" be in the middle of the numbers after cleaning. ^0+ works for START RANGE (10 bytes) But not for the END RANGE (10 bytes) because it will remove the 0s even if the start number is 10 or 100.

[acoder]

An easy way would be to use two rereplaces, one for the first and one for the second number, e.g.

<cfset startRange = REReplace(left(StreetNumber,10),"0+","")>
<cfset endRange = REReplace(right(StreetNumber,11),"^0+([1-9][0-9]+)[OEB]","\1")>
<cfset StreetNumber = startRange & endRange>

[cfmx2008]

You’re awesome!!!!!
It's done exactly the way I wanted. Thank you sooooooo much for giving me your time. I really appreciate it.

[acoder]

No problem, you're welcome as always.