Traversing a Column with Java Snippet Simple

KNIME Analytics Platform
1

Hi,

I need to loop through my data set with a java snippet and create a MeasureValue column.

For example, my data looks like this

PodcastNumber,Location,Group,Metric
1,STL,1,M1
1,STL,0,M1
1,DAL,1,M1
1,DAL,0,M1
1,STL,1,M2
1,STL,0,M2
1,DAL,1,M2
1,DAL,0,M2
2,STL,1,M1
2,STL,0,M1
2,DAL,1,M1
2,DAL,0,M1
2,STL,1,M2
2,STL,0,M2
2,DAL,1,M2
2,DAL,0,M2
…

I need to derive a metric value for each podcast for each location for each group, and for each metric.

Something like this…
Let X = 1;
For (Podcast X in Podcasts)
For (Location X in Locations)
For (Group X in Group)
IF Metric = M1 THEN Value1 ELSE Value 2

Can somebody please help me with this?

2

Hi @cageybee -

If I’m understanding you correctly, don’t you just need to loop over each record in your dataset, as opposed to using a nested loop structure?

If so, you can just use the Rule Engine node, and append a Value column based on your Metric.

3

I can’t use math with Rule Engine and this is the reason why I am using the Java Snippet.

4

Would something like this be helpful?

5

Scott, this is where I’m at right now. I am using an IF ELSE IF ELSE in my snippet. However, it is not very efficient as I have 50 podcasts, 8 locations, 2 user groups, and 9 Metrics.

6

How about a set of look up tables? This depends on exactly how your value is being calculated, but e.g.

Podcast Number, MeasureValue
1, ...
2, ...

And then

Location, MeasureValue
STL, ....
DAL, ....

and so on…
Then, use a series of Joiner nodes to join your data set to the corresponding look-up tables, and finally a Math Formula to sort out your final MeasureValue column from the individual contributions?
Whether this works as a suggestion depends on how that MeasureValue is calculated, but your comment that

suggests something like this might work?

Steve

7

Thanks, Steve, I will try this to see if it will work.