Hi guys,
I see at others platforms like knime (pentaho) a node that can manipulate a list of rules/expressions at once, avoiding to create others nodes lioke a step-by-step for the same subject.
Example:
Step1 - Chance string case
Step2 - Trim spaces
Step3 - replace a string
Step4 - split the value into columns…
all of it in just one node… and you can choose with which field you can use/return or create a new one too…
It some like the rule engine node that you can write lines to each rule, but you select fields, kind of field, manipulation, regular expression or not, pattern… well, a lot of stuffs as print below… It will marvelous if knime could do the same thing or better than that…
As you can see, all this print give more liberty to work with multiples rows/tasks for manipulation or calculation, simplifying the number of steps/nodes inside the workflow.
I like very much knime, but it can be better then now too… some insights for you!
Thanks,
Denis
Hi Denis,
Thanks for your feedback. Maybe the Column Expressions is something for you?
Kind regards,
Alexander
1 Like
Hi @AlexanderFillbrunn ,
It helps, but I need to build a lot of simples cases that could exists as examples pre-defined, to just select options as the given prints, to be more friendly. Each case (string and calculation) have a bulk of tools and the possibility to use regular expression just selecting the field and making a choise (fill the blanks for “dummies”).
But, it helps…
1 Like
Maybe to add, for sure the CE has all of these feature but it always requires the function to be defined manually unless you create a component for it utilizing a bunch of configuration nodes and flow variables.
For example in the Pentaho String Operations node, it’s a matter of just selecting upper/lower for the case, trim type left, right or both, etc.
It’s a low effort, quick solution to perform the most common string operations. Having worked with Pentaho for a few years, the string operations node is about the only thing I’d wish KNIME had.
3 Likes
Yeap, the string manipulation works very fine for this case, I’ve projects that at this node, validate and manipulate more then 20 tasks at once very fine, avoiding a lot of nodes… Json and loops are the best part of knime, winning from pentaho far far away…
But it could be a official node, not a lab node/experimental … with some adjusts.
BR,
Denis
