Tranforming a continous set of data

KNIME Analytics Platform
1

Hey,
I am still new to Knime and I do have a question.
I have file which contains a continues set of data:

measurement time
7960 08:20:03
7960 08:20:03
7960 08:20:03
7960 08:20:03
…
7200 08:20:09
5800 08:20:09
4365 08:20:09
2060 08:20:09
2060 08:20:09
4569 08:20:09
6598 08:20:10
7960 08:20:10
7960 08:20:10
…

What I need to do is to create and new table which looks like the follwing:
time_start time_end measurement(max-min)
08:20:09 08:20:10 5900

I coded it as an excel makro and it worked but the idea is to do the whole data configuration and transforming in Knime without any Java code. Is there any possibily to do that? Attached you’ll find the file with the data. Help is much appreciated!
Thanks and regards
Nico05031728.xlsx (128.2 KB)

2

have you looked at the “pivoting” node already? has some interesting transformation & aggregation capabilities!
–there’s also the unpivot to achieve the opposite :wink:

H

3

Hi Nico,

You can do this very easily with a GroupBy node, no Java need :wink:

See the attached example. After converting to Date&Time format, I use the GroupBy to get the min and max of the timestamp as well as the range of the measurement.

Is this what you were looking for?

Cheers,
Rolandtimeseries_example.knar (76.3 KB)

4

Hi Roland,
thanks so much for your help! That’s pretty close to what I need.I attached an Excel for better understanding. It is created out of the same data set, I recently attached. It’s different from your solution because every time the value in column A drops under 7960/7950 for more then 10 values in a row a new row is created where the max and min in this range is calculated. Is that also possible in knime?

Thanks so much for your help!

Cheers,
Nico

Adhoc-Sensoric.xlsx (8.3 KB)

5

Hi,
we tried a different alternativ and it worked! Thanks so much for all your help!

Cheers
Nico

6

Hey @Nico_95,

do you mind sharing your successful approach? KNIME community relies on sharing - both questions and answers.

Thanks in advance and kind regards,

Patrick.

1 Like