Forum Discussion
Plot Line Chart (Industrial Logging) without Aggregation
- 2 years ago
The Y-axis in a line chart functions as a value axis, inherently performing aggregation, much like a pivot table in Excel.
However, it's essential to note that the aggregation always occurs at the level of each point-to-point interval on the X-axis. For example, if you have data recorded every second with the date and time on the X-axis categorized accordingly, the chosen aggregation method—whether it's sum, average, minimum, or maximum—will consistently display a single value for each interval . This is because the aggregation is accurately filtered to the specific point, resulting in the same value for that particular interval . The X-axis of the graph should be "categorical"
I put on the attached picture graph and table with "don't summarize", you can see that the values are equal:
The issue with the instruments in your case is that you can't show the same time stamp for every occurrence,
X-Axis cannot have duplicates, so you need to filter the graph by instrument or work with a hierarchy
or from instrument to logs and then it would look like this (my country is your instrument) :or from instrument to timestamp and it would look like this:
the file with the hierarchies is attached
If this post helps, then please consider Accepting it as the solution to help the other members find it more quickly
Can i understand it as if the x-axis is so granular that there will be only 1 y-value for each x-value, any aggregation will essentially be the same as the value itself without aggregation?
But powerbi has to cater to the more general case where there may be multiple points per x-axis value, so it forces aggregation by default?
One example where I see this show up is when you have a datetime column on x. If you use the automatic "Date Hierarchy" , there will be multiple points per x-axis.
If you click the dropdown on the field in the Visualizations -> Build visual tab to switch to the original column name, and if your timestamp is granular enough, there will only be 1 point per x-value. If this happens, just ignore the aggregation applied.