Finding smallest common product id

KNIME Extensions
1

Hello

My new challenge is to shorten a big list of our products (round about 120k) by their responsible R&D guy. Well this would be easy with the “group by”-node.
My challenge is now to find the smallest common product id. The product id is of course unique and a 12 to 16 digit/letter combination.

It’s like 1ab2345-6cd78-9ef0. One group can end with at least 6 digits/letters another with 8 or 4.

Is there a node for that? Did somebody such a job in the past? I am open any help.

Thanks in advance.

br, martin

2

Hi,

So you want to find the IDs ending in 4 digits/letters? (Or generally IDs that ends with the least number of digits/letters)

3

Hi
I am not fixed to a certain amount of digits/letters. It should be the most least number of digits/letters, depends on the variance of products and responsible persons.

In my first try I grouped by product family (first 3 digits/letters) and the responsible person. For example, now I 6 row with kindly different product Id’s. 1ab222, 1ab245, 1ab285, 1ab269, 1ab612 and 1ab902. The first four should end after 5 digits/letters and the last two can end after 4 digit letters.
How can I shorten them automatically?

4

Hi Martin!

Not quite sure what are you trying to accomplish. Maybe provide an example input and output and someone might help :wink:

Br,
Ivan

1 Like
5

Here is a method to extract the last portion of the ID after the “-”

Pattern p = Pattern.compile("\\-(?:.(?!\\-))+$");
Matcher m = p.matcher($id$);

if(m.find()) {
  return m.group(0);
}
else {return "";}

image
image678×557 91.8 KB

You could then decide what to do with the results, how to sort them; you might extract the numbers or come up with some logic.

image
image848×466 60.2 KB

kn_example_text_number_value_regex.knwf (18.6 KB)

2 Likes
6

So the first part of the IDs (the digits/letters before the first -) relates to the product family and the responsible person and you want to take last part (the digits/letters after the last -) of the IDs.
If that’s the case then the solution by @mlauber71 is doing what you want and also you can use this expression in a “String Manipulation” node to do that for you:
regexReplace($column1$, ".*(?=-)-", "")

But if this is not what you want I need more explanation.

Best,
Armin

2 Likes
7

Thanks for your answers, but I think this is not exactly what I’m searching for.
It is not only the last part. Please have a loo at this example table:
2019-02-25%2009_31_02-Mappe1%20-%20Excel

The minimum count is 3 digits, like the first column. I hope this makes it clearer.
Oh, just for your information, 0X can also change. :slight_smile:

Edit:
As I investigated my table I realized, that I did not catch all possible groups, because e.g. Person A is also responsible for 0XA21 and Person B for 0XA23. I configured my groupe by node with “last”. :frowning: