Changing Timeseries from 10 minute to hourly

Is there a way to convert my timestamp, from a 10min interval to an hour interval?

Hi @shawn,
welcome to the Community.

I think extracting the date and the hours of your timestamps followed by a groupby node summing the values up based on group conditions date and hour should work.


1 Like

Yep I used the “Create Date&Time Range” node, to edit the interval to hours. But how do you do that on the groupby node? Im sorry, still quite new to knime.

After transforming your timestamps you normally have multiple records for the same interval. Therefore you need to aggregate the data to have an unique record. The aggregatation depends on your readout.
Sometimes you have to sum up the data e.g. if you have counting information. Sometimes you have to build an average with in addition descriptive stats. And this can be done with the group by node.

So initially my data was this, which is of a 10min interval range.

I then converted to a 1hr range image

If i got you right you now have 2 tables. One with the data and a second with your inteval and you now have the problem to bring both together, right?

Yep so will be using joiner to join them up. But my data set is over the course of 2 months, so is there a way I can simplify it to just 1 day? Prob get the average of all the days in the 2 months?

I would solve it only based on your source table.
transforming your Timestamp with the Date&Time to String node with format yyyy-MM-dd HH:00 and then convert it back to a Timestamp with String to Date&Time node using the same format.

This converts your Timestamp e.g. from 2016-01-11T17:10 to 2016-01-11T17:00. Now you have multiple records for each hour with the same timestamp which you have to aggrgate.

1 Like

Actually what is the purpose of using Date&Time to string and then String to Date&Time?

removing the minutes of your timestamp and only keep the date and hour information of your source data.

Alright understand Thanks

However I realise that when I do that, I am only left with my timestamp of interval 1 hr, in a way all my other observations are lost. I tried using concatenate but it there is a huge difference in row number which causes an issue

Don’t replace (overwrite) your origin Timastamp column and you have both information for further manimpulations available.

Hi @shawn,

also check out the aggregation granularity component [1]. For hourly aggregation you can use it out of the box. It basically creates a new column that is used to group on (e.g. “2021-12-24 15”) in a group by node (check out the Table Creator node insight the component for more information).

It is part of the time series analysis collection of Corey, Daniele and Maarit [2]

Best regards,

[1] Aggregation Granularity – KNIME Hub
[2] Time Series Analysis with Components | KNIME

1 Like

This topic was automatically closed 182 days after the last reply. New replies are no longer allowed.