Black Lives Matter. Please consider donating to Black Girls Code today.

Plotly.js removes rows from a table type trace inserted in a subplot

Seven days ago I uploaded to Plotly cloud a plot like in the attached image.

Meanwhile it turned into this weird one

The table y-domain is [0,1], but plotly.js removes the last 16 table rows leaving that gap.

@etienne This is the definition of my table:

table_trace1=dict(type = 'table',
              domain=dict(x= [0.0, 0.5],
                         y= [0, 1.0]),
             columnwidth= [30]+[ 33, 35, 33],
             columnorder=[0, 1, 2, 3, 4],
             header = dict(height = 50,
                           values = [['<b>Date</b>'],['<b>Hash Rate, TH/sec</b>'], 
                                     ['<b>Mining revenue</b>'], ['<b>Transaction fees</b>']], 
                           line = dict(color='rgb(50,50,50)'),
                           align = ['left']*5,
                           font = dict(color=['rgb(45,45,45)']*5, size=14),
                           fill = dict( color = '#d562be' )#fill color for header
             cells = dict(values = [df['Date'][-20:], df['Hash-rate'][-20:], df['Mining-revenue-USD'][-20:],
                          line = dict(color='#506784'),
                          align = ['left']*5,
                          font = dict(color=['rgb(40,40,40)']*5, size=12),
                          format = [None]+ [",.2f"]*2+[',.4f'],  
                          prefix = [None]*2+['$', u'\u20BF'],
                          height = 27,
                          fill = dict( color = ['rgb(235,193, 238', 'rgba(228, 222,249, 0.65)'] )

I followed .

Hmm. I don’t think this is a plotly.js problem.

I know we weren’t saving table traces correctly in our database when we first introduced table traces.

Does the problem happen on newly-created tables as well?

@etienne I re-ran the corresponding notebook and uploaded it here, with both offline and online plots
The offline plot is OK (plotly 2.1.0), but the online one has that gap.

@monfera got an idea why this could be happening?

Maybe could be the culprit?

There’s no support yet for table on Vera has been working on it. It looks like the data are transposed for certain things, the original data has 4 columns and this shows 4 rows; also, reading numbers in a specific row reveal that they should rather be in a column. I heard that table orientation causes challenges that are being addressed in the upcoming PR for integration - ie. the integration work has deeper aspects than just providing the UI elements for building a table plot online.