Skip to main content
Announcements
Qlik Introduces a New Era of Visualization! READ ALL ABOUT IT
cancel
Showing results for 
Search instead for 
Did you mean: 
ischmidt20
Contributor
Contributor

How to ensure line / area chart displays ALL points

Hello,

I have daily data across ~40 individuals spanning 4-5 years. On each day, each individual is in one of 6 possible statuses. I would like to plot the total number of individuals in each status in each day as a stacked line chart. However, when I do that, I see a "currently showing a limited data set" warning, and the chart looks like this:

ischmidt20_0-1712612489357.png

This does not make sense to me as the top line should be flat, as the number of visuals is not changing and there is no missing data. If I change "show missing values" setting to "show as gaps," it looks even worse:

ischmidt20_1-1712612576887.png

In an attempt to compromise, I created a measure to roll up the daily data to the weekly level, thus reducing the number of data points to ~14% of what it was before. The "currently showing a limited dataset" warning goes away, but clearly not all points are being shown. Plus, I don't want weekly data anyway, I want to keep the data at the daily level.

ischmidt20_2-1712612672720.png

My colleague figured out a way to ensure the top line is flat, effectively defining 6 different measures, one for each status. He also managed to get rid of the "limited dataset" warning:

ischmidt20_3-1712612759560.png

However, when I filter to one individual, I get these weird slanted lines, when really they should be very close to vertical/a step function (considering the values will go from 0 to 1 or 1 to 0 in a single day, and I am showing a whole year):

ischmidt20_4-1712612863203.png

So clearly, this is not showing all data points either. This is confirmed by the fact that when I select about a year's worth of dates in the line chart with the lasso tool, only ~15 dates will appear as selected (when really it should be ~365).

In all of these cases, all of the data is available as when I download the dataset from the visual, it is a complete file. How can I get Qlik to show every single data point?

To give you some idea of what I am looking for, here is how the chart looks when I plot with matplotlib in Python:

ischmidt20_6-1712613391775.png

Unfortunately I cannot share the Qlik file or the underlying data as it is business-sensitive.

Labels (1)
2 Replies
Lisa_P
Employee
Employee

There are max no of points limitations in line charts:

  • Max visible points: Set the maximum number of points that will be displayed. The default is 2,000. The maximum is 50,000. If you set a number less than 1,000, the line chart will behave as if the maximum is 1,000 visible points. The actual maximum number of data points in the chart is affected by the distribution of the data, and could be lower than the value you configure with this setting. When there are more data points than the value you have set, data points are neither displayed, nor included in selections made in the chart.

  • To get rid of the angle when you are looking at limited values, you can change the measure expression to use null for zeros eg IF(count (distinct Employees)>0, count(distinct Employees), null())
ischmidt20
Contributor
Contributor
Author

Thank you Lisa. Unfortunately, my system uses Qlik Sense from Nov 2022, and it looks like this "Max visible points" feature was implemented after. I have asked about updating to a more recent version, but is there any other alternative?

Also, the fact that I see this behavior even when filtering to just one individual (< 2000 data points), leads me to believe that there is something else going on: "The actual maximum number of data points in the chart is affected by the distribution of the data." Could you elaborate more on this?

Thank you for the suggestion about converting zero values to Nulls. Unfortunately, this does not resolve my issue. With showing missing data as gaps:

ischmidt20_0-1712672456635.png

As connections:

ischmidt20_1-1712672481611.png

And "as zeros" is the same slanted lines as before. The closest is the "show as gaps," but really that is wrong because, well, there is a gap where there should be data.