flow variable with String manipulation . . .

[sanderlenselink]

Hi there,

I’m struggling with some data wrangling . . .

I have dates for a number of companies (not every company has the same dates).
And I have a start date that varies by company

What I’m looking for . . .

• all dates from the start date get a “1”
• and all date older than the start date get a “0”

I’m thinking of doing something with flow variables . . . but I don’t see how

Can anyone get me started?

Many THNX in advance

BREAKOUT_knimeforum_20230802.knwf (17.8 KB)

[Daniel_Weikert]

Hi
have you considered converting start date to a variable and use it in the Datetime difference node to calculate the difference in days. Then sth like rule engine if the difference is positive 1 else 0 ?
br

[sanderlenselink]

Hi @Daniel_Weikert

yes, I toyed with that idea but in practice I see a huge complexity with the dataset. There are 100+ companies and 100+ different start dates. And that grows every week with new 5 companies and 5 new start dates.

To keep it simple I thought about using the start dates as flow variables. But I don’t see how

With this complexity, would you stick to “start dates as variables” or switch to another approach (e.g. flow variables or looping).

THNX for your suggestions

-SL

[ArjenEX]

I’d say @Daniel_Weikert is on the right track here. Something like this could do the trick:

I opt for an one-node solution using the CE node but that’s personal preference.

if (rowId().equals("START")) {
    null
} else if (column("comp1") >= variable("comp1")) {
    1
} else {
    0
}

If your dataset is that large I think you can’t escape looping and using the flow variables from the loop as input (like CurrentColumn).

WF:
BREAKOUT_knimeforum_20230803 V2.knwf (27.7 KB)

Hope this helps!

[sanderlenselink]

Hey @Daniel_Weikert & @ArjenEX

you both thnx for your feedback and now I was able to create the flow I needed.

I slightly adapted my requirements:

• if START equals a date it generates the value “9” instead of issing
• if a date is missing the end situation gives a missing ("? " )
  This to avoid that “equals” and “missing” are mixed.

Just a litte question to @ArjenEX . . .
in the Column Expression node you use “.equals”
if (column(“ColumnValues”).equals(column(“START”)))

I would use “==” but that doesn’t work . . . why ???
if (column(“ColumnValues”) == (column(“START”)))

regards

BREAKOUT_knimeforum_20230806 V3.knwf (27.2 KB)

[ArjenEX]

My recommendation would be to always use equals if you are looking for a particular text string. More details about what’s best to use when:

Difference between comparing String using == and .equals() method in Java -...