Line plot with category_orders draws the line in the wrong order

[ghost]

I am trying to plot a pandas DataFrame with a column that contains the month as a string:

df = pd.DataFrame({"month": ["feb", "mar", "jan", "apr"], "value": [20, 3, 10, 4]})
df

	month	value
0	feb	20
1	mar	3
2	jan	10
3	apr	4

If I just feed this to px.line the x-axis is in alphabetical order (feb, mar, jan, apr):

px.line(df, x="month", y="value")

Luckily px.line has a category_orders argument:

px.line(df, x="month", y="value", category_orders={"month": ["jan", "feb", "mar", "apr"]})

It changes the x-axis order, but it draws the line in the original order: from feb to mar to jan to apr

Should I call the px.line in another way when I want to use that argument? Because the current behaviour seems incorrect.

I can work around it by first sorting my DataFrame and then plotting, but it would be more convenient if the category_orders argument could be able to take care of it.

monthdict = {"jan": 1, "feb": 2, "mar": 3, "apr": 4}
df.sort_values(by="month", inplace=True, key=lambda x: x.map(monthdict))
px.line(df, x="month", y="value")

I founde #3645, but that seems to be in combination with the facet_col argument. I am experiencing this issue with a "plain" line.

Version of plotly is 5.10.0.

[nicolaskruchten]

The current behaviour is as designed: lines connect points in the order they are given in the input data frame, and the category_orders controls the axis order only, independent of the data frame order.

[ghost]

Ok, that's odd, but consistent I guess. In that case this ticket can be closed.