Hi,
I am a newbie in KNIME and I have an issue on extracting a number to a string. The number that I need are in different locations. Sample are below:
81727904HG ASDHASD 275 EE
Resource - 70331103
Ma 922803694FF400g*24
Output I need are:
81727904
70331103
922803694
Sorry last item should be:
input: Ma 92280369FF400g*24
output: 92280369
Welcome to the forum @AlyKnime .
Do you have more examples of strings with numbers you’re trying to extract? From what you’ve provided it looks like a regex string that matches numbers with more than 4 digits would suffice.
I used the Regex Extractor node (part of the Palladian collection), with the string \d{4,20} which finds numbers with 4 to 20 digits. If you have more complicated data then this may need to be modified.
Hi @elsamuel
Sample.xlsx (270.2 KB)
I have attached the sample of strings. actually there are rows that do not have a “code” that i need.
The file has two columns, a short text and a code. in the short text, there are items that includes the codes.
BTW I am not really familiar with the Regex Extractor 
Hi @AlyKnime,
Here is an example for you, I extracted the codes with both String Manipulation node and Regex Extractor node, use each one you prefer.
For installing and using Palladian nodes follow this instruction : Palladian [Product] — NodePit
GL,
Mehrdad
Hi @AlyKnime , you should first confirm if the logic proposed by @elsamuel makes sense, that is “finds numbers with 4 to 20 digits”. You can see why he’s doing greater than 3 digits, because for example, from “81727904HG ASDHASD 275 EE”, if you extract the numbers without any restrictions, you would get 81727904 and 275, similarly for “Ma 92280369FF400g*24”, you will get 92280369, 400 and 24.
So, do the numbers you are targeting always have a specific range of digits (minimum and maximum)? Are they always between 4 and 20 digits?
Hi @mehrdad_bgh ,
I did try this, but encounter an error on the output for string, 81715738180ML HE CND BAMBOO. I had an output of 17157381 instead of 81715738.
Hi @AlyKnime , then from @elsamuel regex, you only need \d{8}
Be advised though that if you have data that may contain 2 sets of 8 digits, then both sets would be return. For example “81727904HG ASDHASD 27527527 EE” would return 81727904 and 27527527.
Hi @bruno29a, is it via the regex extractor or in string manipulation? Cause what I am using is the string manipulation. Thanks!
Hi @AlyKnime , I put something together.
It looks like this:
Input data is:
Output data is:
And here is the workflow: Extract 8 digit numbers from string.knwf (6.7 KB)
It’s done via Regex Extractor as it’s straight forward for what you need - extract the numbers. You cannot do this in 1 operation with String manipulation.
Is Regex Extractor an issue because you don’t have Palladian extensions?
EDIT: Correction: After viewing what @mehrdad_bgh has done, it looks like it can be done in 1 operation with String manipulation with the regexReplace. @AlyKnime , the regex probably needs to be adjusted. I’m not too strong with regex, so I can’t really help for the expression.
Yes that’s right I don’t have the Palladian extensions. and wondering if I can use the string manipulation instead 
Hi @mehrdad_bgh , just wanted to ask why do you think it captures the 17157381 instead of 81715738? when the “8” comes first than “1”? Just confuse on the logic Sorry and thanks for your help!
Yes, it looks like it takes 8 digits from the right - if you change from 8 to 6 for example, you can see this behaviour for all of the input data.
