如何从 Excel 图表图例中删除系列 - Python Openpyxl

Question

I am creating a custom chart style using a ScatterChart from openpyxl v3.0.7. I want to show a legend only containing certain series within the chart, as the rest are only needed for structuring the underlying chart and not pertinent to the data relevant to a standard consumer.

Here's an example of the chart code:

import openpyxl as xl

wb = xl.workbook.Workbook()
sh = wb['Sheet']
data = [(1, 2), (2, 3)]
for row in data:
  sh.append(row)

hideCnt = 0
serName = 'Test'
xVals = xl.chart.Reference(sh, min_row=1, min_col=1, max_row=2)
yVals = xl.chart.Reference(sh, min_row=1, min_col=2, max_row=2)
ser = xl.chart.Series(yVals, xVals, title=serName)
hideCnt += 1

ch = xl.chart.ScatterChart()
ch.title = 'Chart-1'
ch.series.append(ser)
ch.legend.legendPos = 'b'

entries = []
for i in range(hideCnt):
  entry = xl.chart.legend.LegendEntry(idx=i, delete=1)
  entries.append(entry)
ch.legend.legendEntry = entries

sh.add_chart(ch, 'A3')
wb.save(filename = 'test.xlsx')

This generates an excel output as such: Chart Example w/ Legend (series remain)

The series remains in the Legend of the chart, despite the delete=1 tag on the legendEntry for series of idx=0 . I have looked elsewhere and found no solution that works for deleting these series from the legend using openpyxl. Per a similar question, How to remove a legend name in openpyxl , I was led to this post: LegendEntry does not contain delete element when serialised , last updated a year ago.

Here is the resultant XML in 'test.xlsx\xl\charts\chart1.xml' from my code:

<chartSpace xmlns:a="http://schemas.openxmlformats.org/drawingml/2006/main" xmlns="http://schemas.openxmlformats.org/drawingml/2006/chart">
 <chart>
  <title>
   <tx>
    <rich>
     <a:bodyPr/>
     <a:p>
      <a:pPr>
       <a:defRPr/>
      </a:pPr>
      <a:r>
       <a:t>Chart-1</a:t>
      </a:r>
     </a:p>
    </rich>
   </tx>
  </title>
  <plotArea>
   <scatterChart>
    <ser>
     <idx val="0"/>
     <order val="0"/>
     <tx>
      <v>Test</v>
     </tx>
     <spPr>
      <a:ln>
       <a:prstDash val="solid"/>
      </a:ln>
     </spPr>
     <marker>
      <symbol val="none"/>
      <spPr>
       <a:ln>
        <a:prstDash val="solid"/>
       </a:ln>
      </spPr>
     </marker>
     <xVal>
      <numRef>
       <f>'Sheet'!$A$1:$A$2</f>
      </numRef>
     </xVal>
     <yVal>
      <numRef>
       <f>'Sheet'!$B$1:$B$2</f>
      </numRef>
     </yVal>
    </ser>
    <axId val="10"/>
    <axId val="20"/>
   </scatterChart>
   <valAx>
    <axId val="10"/>
    <scaling>
     <orientation val="minMax"/>
    </scaling>
    <axPos val="l"/>
    <majorGridlines/>
    <majorTickMark val="none"/>
    <minorTickMark val="none"/>
    <crossAx val="20"/>
   </valAx>
   <valAx>
    <axId val="20"/>
    <scaling>
     <orientation val="minMax"/>
    </scaling>
    <axPos val="l"/>
    <majorGridlines/>
    <majorTickMark val="none"/> 
    <minorTickMark val="none"/>
    <crossAx val="10"/>
   </valAx>
  </plotArea>
  <legend>
   <legendPos val="b"/>
   <legendEntry>
    <idx val="0"/>
    <delete val="1"/>
   </legendEntry>
  </legend>
  <plotVisOnly val="1"/>
  <dispBlanksAs val="gap"/>
 </chart>
</chartSpace>

The resultant XML does actually have the correct delete tags associated with the legendEntry at idx=0, which correlates to the correct chart series, 'Test', w/ series <idx val="0"/> . However, the end result is the same. If I save the file after opening in Excel then the XML changes and all of the legendEntry tags are removed.

XML legend excerpt after saving from Excel:

<c:legend>
 <c:legendPos val="b"/>
 <c:overlay val="0"/>
</c:legend>

In addition to removing the legendEntry elements, there seems to be a prefix added to all elements that were previous left without a namespace (from the original openpyxl output).

http://schemas.openxmlformats.org/drawingml/2006/chart

I have attempted a workaround by loading the modified workbook w/ adjusted namespaces into openpyxl and changing the charts after the fact, but I have had no luck accessing existing charts from a saved workbook.

Does this seem related to the namespace difference between the Excel version I'm running and the output generated by openpyxl? If so, is there a way for me to modify the namespaces to adjust accordingly? If not, what am I doing wrong?

Answer 1

It looks like you may need to add a c:numCache node (under c:numRef node) to get the legend entry deletion to be recognized. Based on your xml above, that probably will be numCache node under numRef.

I ran into the same issue with Microsoft's OpenXml library for .NET. So this issue applies to more than just openpqxl but to the OpenXml format in general.