Scatter chart - two data points not in date order

[Richard_T]

I have an issue with two data points on my XY scatter chart which are not in date order as shown in Fig.1. In this chart, I am trending the output of a manufacturing process ("Crystallinity", Y-axis) against the date the process took place ("Spray Date", X-axis). A legend is added to denote the batch of raw material used. As you can see, I have two data points with a spray date of 10 April 2022 which is displayed between 17 March 2022 and 18 March 2022 but I don't understand why it's out of sequence.

I've provided the background of my data and the queries used further below. Can anyone point out if I'm missing something obvious or is this a bug?

Fig.1

Background information on data and queries

The settings for the X-axis are set to "Don't summarise" and set by "Spray Date" (instead of "Date Hierarchy") as shown in Fig.2.

The data source is an Excel sheet on a network drive where the data is entered on the date of spraying (i.e. chronological order). I double checked the Excel sheet and there's nothing different about the date format of these entries compared to the rows above/below it. During the import of data to Power BI, I added a couple of queries to transform the data which includes the assignment of the date format for "Spray Date". All queries pertaining to "Spray Date" are shown in Fig.3.

After all queries are complete, I am left with the data shown in Fig.4. The two data points in question are listed in chronological order as per the Excel sheet and nothing looks out of the ordinary for these rows compared to the others. No sorting has been applied to the tabulated data.

Fig.2

Fig.3

Fig.4

ETA: Fix grammar

[mahenkj2]

Hi @Richard_T ,

If possible, can you please share data in table format, just to easily copy and test at our end? Please remove sensitive data, only sample data should be enough.

[Richard_T]

Hi @mahenkj2, see the post from FarhanAhmed below. It looks like my data has been sub-sorted by date within each category that is listed in my legend. 

https://community.powerbi.com/t5/Desktop/Scatter-chart-two-data-points-not-in-date-order/m-p/2560825...

As requested, here is some sample data between 01 March 2022 and 30 April 2022 so you should be able to recreate the issue I had with the data around 10 April 2022. Let me know how you get on.

Spray Date (x-axis)	Powder Batch No (legend)	Crystallinity (y-axis)
01 March 2022	1500484	58.2
02 March 2022	1500484	59.3
03 March 2022	1500484	61
04 March 2022	1500484	60.5
05 March 2022	1500484	60.2
06 March 2022	1500484	58
07 March 2022	1500484	62.8
08 March 2022	1500484	62.4
09 March 2022	1500484	60
10 March 2022	1500484	62.5
11 March 2022	1500484	65.9
12 March 2022	1500484	62.4
13 March 2022	1500484	67.6
14 March 2022	1500484	65.9
15 March 2022	1500484	61.8
16 March 2022	1500484	58.2
17 March 2022	1500484	54.4
17 March 2022	1506369	54.5
17 March 2022	1506369	59.6
18 March 2022	1506369	59
19 March 2022	1506369	65.4
20 March 2022	1506369	63.36
21 March 2022	1506369	62.62
22 March 2022	1506369	59.79
23 March 2022	1506369	57.2
24 March 2022	1506369	58.89
25 March 2022	1506369	59.86
26 March 2022	1506369	58.68
27 March 2022	1506369	57.2
28 March 2022	1506369	56.3
29 March 2022	1506369	52.19
29 March 2022	1506369	50.44
29 March 2022	1506369	53.62
30 March 2022	1506369	59.97
31 March 2022	1506369	62.67
01 April 2022	1506369	58.54
02 April 2022	1506369	60.71
03 April 2022	1506369	60.73
04 April 2022	1506369	57.15
05 April 2022	1506369	61.28
05 April 2022	1506369	62.15
06 April 2022	1506369	59.19
07 April 2022	1506369	63.87
08 April 2022	1506369	65.37
09 April 2022	1506369	62.89
10 April 2022	1506369	62.45
10 April 2022	1506364	61.95
11 April 2022	1506369	58.55
11 April 2022	1506369	61.78
11 April 2022	1506369	69.94
11 April 2022	1506369	64.53
12 April 2022	1506369	64.61
13 April 2022	1506369	60.96
14 April 2022	1506369	61.93
15 April 2022	1506369	61.43
16 April 2022	1506369	61.82
17 April 2022	1506369	59.21
18 April 2022	1506369	58.02
19 April 2022	1506369	62.94
20 April 2022	1506369	57.67
21 April 2022	1514289	57.89
22 April 2022	1514289	50.62
22 April 2022	1514289	60.45
23 April 2022	1514289	59.66
24 April 2022	1514289	63.6
25 April 2022	1514289	62.59
26 April 2022	1514289	60.36
27 April 2022	1514289	56.49
27 April 2022	1514289	57.37
28 April 2022	1514289	66.5
29 April 2022	1514289	67.71
30 April 2022	1514289	61.3

[mahenkj2 · Accepted solution]

Hi @Richard_T ,

I have tested your data and could replicate the problem and also understand the grouping by date and legend together and then sorting.

Since last 2 days, I am trying sorting of x axis date with an index column, but somehow it does not allow date.

One more workaround is to use line chart in place of scatter chart, and not use the line color but just the marker color, and sort by date. It works fine even with legend, but you need to specify legend colors manually while publishing the report.

Hope it helps.

[Richard_T]

Perfect, this is exactly what I was looking for! I didn't think to use a line chart as I thought scatterplot would be the right tool.

I just changed the chart type and I didn't even have to move any of the fields. See below picture, which now shows the two data points correctly at 10 April 2022 instead of sitting within March.

[FarhanAhmed]

Did you checked your "Sort By" in Visual

[Richard_T]

Hi Farhan. Originally, I had not applied any sorting to the chart. I added a sort by "Spray Date" and set to "Sort ascending" but the data is still out of sequence as shown in this new screenshot.

[FarhanAhmed]

If you see that it is grouping the data based on the legends. then it is applying the sorting.

you may want to use another visual if you want your x-axis to be sorted by Date without grouping. You may try out Dual Axis Scatter Chart 

[Richard_T]

I can confirm that removing the legend puts the data back chronological order as shown in the picture below.

Whilst this solves the problem, it's not a workable solution for me. I need to easily see when a powder batch change occurs and correlate this with any change in result on Y-axis. Tooltips is too slow to check for every point so the legend/colour coding data points is the most logical solution.

Is there really no way to add a legend and not subgroup the data? 

[FarhanAhmed]

Hi @Richard_T 

There is one work around that you can check if possible in your scenario

you can assign color to different category using DAX 

Color = SWITCH('Table'[Product],
"A","Blue",
"B","Green",
"C","Yellow",
"D","Gray",
"Brown")

Then use Marker conditional color expression to Field Value

Then you will get results like below sorted by date

but you will not get the Legends as what color represtenting what data.

may be a table right next to visual something like below to give view like legends. similar conditional formatting apply on the table 

see if this works for you.

[Coriel-11]

This worked for me. Thanks

Matt

[FarhanAhmed]

Try using this visual from Store