mplfinance: 'secondary_y=True' not working. Not plot multiple lines/candles due to very wide Y-Axis data ranges

Question

I want to compare price movement of two scrips with Stock Indices in one chart object (single panel) using mplfinance in python 3.11.3.

Tried code from snippet: https://github.com/matplotlib/mplfinance/blob/master/examples/addplot.ipynb

I have data in 3 excel csv files that are to be plotted (line or candle) in one chart:

Indeces data

Data of scrip1

Data of scrip2

Comparison of two tables is generating chart below:

Chart comparing data in two tables

I tried the mplfinance method given in

Plot multiple mplfinance plots sharing x axis

This method exits the console.

Then, the following code has been tried:

import mplfinance as mpf
import pandas as pd

idc = pd.read_csv(r'D:\NUVAMA_STOCK_DATA\CHARTING-BANKNIFTY-25 MAY 2023\Nifty Bank_03March2023-25 MAY 2023_CHART.csv')
idc['Date'] = pd.to_datetime(idc['Date'], format = "mixed")
idc.set_index('Date', inplace=True)
intraday = idc
intraday.index.name = 'Date'
intraday.shape
intraday.head(3)
intraday.tail(3)

iday = intraday.loc['22-05-2023  09:16:00':'25-05-2023  15:29:00',:]


ce = pd.read_csv(r'D:\NUVAMA_STOCK_DATA\CHARTING-BANKNIFTY-25 MAY 2023\BANKNIFTY 25MAY 2023 CE 46500_CHART.csv')
ce['Date'] = pd.to_datetime(ce['Date'], format = "mixed")
ce.set_index('Date', inplace=True)
intraday2 = ce
intraday2.index.name = 'Date'
intraday2.shape
intraday2.head(3)
intraday2.tail(3)

iday2 = intraday2.loc['22-05-2023  09:16:00':'25-05-2023  15:29:00',:]



pe = pd.read_csv(r'D:\NUVAMA_STOCK_DATA\CHARTING-BANKNIFTY-25 MAY 2023\BANKNIFTY 25 May 2023 PE 43600_CHART.csv')
pe['Date'] = pd.to_datetime(pe['Date'], format = "mixed")
pe.set_index('Date', inplace=True)
intraday1 = pe
intraday1.index.name = 'Date'
intraday1.shape
intraday1.head(3)
intraday1.tail(3)

iday1 = intraday1.loc['22-05-2023  09:16:00':'25-05-2023  15:29:00',:]

ap = [mpf.make_addplot(iday1, type='line', y_on_right=True), mpf.make_addplot(iday2, type='candle', y_on_right=True),]

mpf.plot(
    iday,
    volume=True,
    title='BANKNIFTY&PE 43600', ylabel='Price ($)', ylabel_lower='Volumes', #figratio=(12,4),
    type="line",
    style="sas",
    addplot=ap, 
    )

This code gives following error message:

"D:\PYTHON_3.11.3\Lib\site-packages\mplfinance\plotting.py:694: UserWarning: Attempting to set identical low and high ylims makes transformation singular; automatically expanding.
  volumeAxes.set_ylim(vymin,vymax)
Traceback (most recent call last):
  File "D:\PyProgFiles\ChatGPT\CS_Mpf\TEST_MULTI_BNPECE_StackOverFlow.py", line 44, in <module>
    mpf.plot(
  File "D:\PYTHON_3.11.3\Lib\site-packages\mplfinance\plotting.py", line 780, in plot
    ax = _addplot_columns(panid,panels,ydata,apdict,xdates,config)
         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "D:\PYTHON_3.11.3\Lib\site-packages\mplfinance\plotting.py", line 1075, in _addplot_columns
    ymhi = math.log(max(math.fabs(np.nanmax(yd)),1e-7),10)
                                  ^^^^^^^^^^^^^
  File "<__array_function__ internals>", line 200, in nanmax
  File "D:\PYTHON_3.11.3\Lib\site-packages\numpy\lib\nanfunctions.py", line 483, in nanmax
    res = np.amax(a, axis=axis, out=out, **kwargs)
          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "<__array_function__ internals>", line 200, in amax
  File "D:\PYTHON_3.11.3\Lib\site-packages\numpy\core\fromnumeric.py", line 2820, in amax
    return _wrapreduction(a, np.maximum, 'max', axis, None, out,
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "D:\PYTHON_3.11.3\Lib\site-packages\numpy\core\fromnumeric.py", line 86, in _wrapreduction
    return ufunc.reduce(obj, axis, dtype, out, **passkwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
ValueError: zero-size array to reduction operation maximum which has no identity"

Unable to understand what's going wrong.

DATA:

BANKNIFTY 25MAY 2023 CE 46500_CHART.csv

BANKNIFTY 25 May 2023 PE 43600_CHART

Nifty Bank_03March2023-25 MAY 2023_CHART.csv

Look carefully at the error message. It says `ValueError: zero-size array to reduction operation maximum which has no identity"` ... `"zero-size array"`??? inside an addplot related function? Have you checked and confirmed that `len(iday) == len(iday1) > 1` ?? — Daniel Goldfarb, May 28 '23 at 15:56
grateful for your kind response. Yes checked. len(iday) = 1499, len(iday1) = 1499, len(iday2) = 1499. — DrSaaie K B, May 30 '23 at 07:36
inserted volumeAxes.set_ylim(vymin=1000, vymax=3000000) to try if it works. Its giving syntax error. — DrSaaie K B, May 30 '23 at 07:54
**Please post the actual data (csv files) where they can be accessed so that others can reproduce your problem.** — Daniel Goldfarb, May 30 '23 at 10:38
Not able to upload csv files FROM COMMENTS. I have created a fresh question at https://stackoverflow.com/questions/76367267/csv-data-upload-to-stack-overflow ; from where the data files can be downloaded or accessed. Creation of Google sheet links took too much of time. Hence the delay. Apologies. — DrSaaie K B, May 30 '23 at 17:36
Just fyi, given the above code, and the data files, the code you have provided fails immediately on line `idc['Date'] = pd.to_datetime(idc['Date'], format = "mixed")` with the following error: `ValueError: time data '03-04-2023 09:16' does not match format 'mixed' (match)` ... likely I can work around this, none-the-less be aware. Changing your read_csv to `read_csv(filename, index_col=0, parse_dates=True)` works fine. No need for `pd.to_datetime()` nor `set_index` — Daniel Goldfarb, May 30 '23 at 18:51
I am now able to reproduce the error. Hope to have an answer for you soon. — Daniel Goldfarb, May 30 '23 at 19:00

score 0 · Answer 1 · answered May 30 '23 at 19:29

0

The problem is that the data file "BANKNIFTY 25 May 2023 PE 43600_CHART.csv" contains extra columns at the right that are completely null; note the extra commas on the right:

!head 'BANKNIFTY25MAY2023_43600.csv'
Date,Open,High,Low,Close,Volume,,
24-04-2023 09:16,953.09998,953.09998,953.09998,953.09998,0,,
24-04-2023 09:17,953.09998,953.09998,953.09998,953.09998,0,,
24-04-2023 09:18,953.09998,953.09998,953.09998,953.09998,0,,
24-04-2023 09:19,953.09998,953.09998,953.09998,953.09998,0,,
24-04-2023 09:20,953.09998,953.09998,953.09998,953.09998,0,,
24-04-2023 09:21,953.09998,953.09998,953.09998,953.09998,0,,
24-04-2023 09:22,953.09998,953.09998,953.09998,953.09998,0,,
24-04-2023 09:23,953.09998,953.09998,953.09998,953.09998,0,,
24-04-2023 09:24,953.09998,953.09998,953.09998,953.09998,0,,
!tail 'BANKNIFTY25MAY2023_43600.csv'
25-05-2023 15:22,0.1,0.15,0.1,0.1,1328475,,
25-05-2023 15:23,0.15,0.15,0.1,0.15,876225,,
25-05-2023 15:24,0.15,0.15,0.1,0.1,842600,,
25-05-2023 15:25,0.15,0.15,0.05,0.05,846850,,
25-05-2023 15:26,0.1,0.1,0.05,0.05,1129650,,
25-05-2023 15:27,0.05,0.1,0.05,0.05,471825,,
25-05-2023 15:28,0.05,0.05,0.05,0.05,145525,,
25-05-2023 15:29,0.05,0.05,0.05,0.05,108575,,
25-05-2023 15:30,0.05,0.05,0.05,0.05,71075,,
25-05-2023 15:30,0.05,0.05,0.05,0.05,0,,1501

When you pass an entire dataframe (instead of an array, or series) into mpf.make_addplot(), then mpf.plot() will attempt to plot every column in the dataframe. This is the line of code where you are doing that: ap = [mpf.make_addplot(iday1, type='line', y_on_right=True), ...

With each column that mpf.plot() attempts to plot, it will also attempt to find the maximum and minimum non-NaN values in the column in order to determine the y-axis scale. However for these last to columns there are zero non-NaN values and so it is impossible to determine the min and max.

One possible solution is to get rid of those columns with a line of code something like this:

iday1 = iday1[['Open','High','Low','Close','Volume']]

If you add the above one line of code before calling mpf.make_addplot() then your code will work and the resulting plot will look like this:

answered May 30 '23 at 19:29

Daniel Goldfarb

6,937
5
29
61

1. With due respect, There is something missing. In the plot generated by you, kindly see the thin red line below the two colored plots. That red line is representing the price data in BANKNIFTY 25MAY 2023 CE 46500_CHART.csv (Price range : Max=105.45, Min=0.05. It has not been correctly plotted in the chart generated by your good self. 2. In ORDER TO BRING CLARITY in the chart, I reduced data range using 'iday1 = intraday1.loc['25-05-2023 13:16:00':'25-05-2023 15:29:00',:]'. Pl see https://docs.google.com/spreadsheets/d/1ZgCzJcO-awrTQojVKv6KQLbpTyTyeZGYDr8fnUjShsI/edit#gid=0 Is this a bug? – DrSaaie K B May 31 '23 at 07:45
Confronting several errors. When 'volume=True,' is inserted in 'ap = [mpf.make_addplot(iday1, type='candle', volume=True, ylabel='Price (₹)', y_on_right=True), mpf.make_addplot(iday2, type='line', y_on_right=True)]', it is giving error : KeyError: 'Unrecognized kwarg="volume". Similar error message is given when 'colors=['g','r'],' is inserted. More refining is needed? I am missing something. – DrSaaie K B May 31 '23 at 07:53
Regarding your comment '0' Null values: "The problem is that the data file "BANKNIFTY 25 May 2023 PE 43600_CHART.csv" contains extra columns at the right that are completely null; note the extra commas on the right:", kindly note that there is only one column after OHLC and that is Volumes. The cells containing 0 values were out side the plot range. See the command "iday = intraday.loc['22-05-2023 09:16:00':'25-05-2023 15:29:00',:]" in my original post – DrSaaie K B May 31 '23 at 08:02
When ' volume=True,' is inserted, in "mpf.plot( iday1, title='BANKNIFTY&PE 43600', ylabel='Price (₹)', ylabel_lower='Volumes 10*3', #figratio=(12,4), type="line", volume=True, style="sas", addplot=ap, show_nontrading=True )" it is working. Please see the chart at https://docs.google.com/spreadsheets/d/18K2yR24FgZWIcWlMsTjMaxLiWTVWPKibsVEgkcEZt48/edit#gid=0 How to plot volumes in 2 panels (from different [2] data tables) Confused! – DrSaaie K B May 31 '23 at 08:20
Please excuse me if I am being presumptive, but it appears to me based on your comments that your knowlege of how to ***write and debug*** code is extremely limited if present at all. Instead of using ChatGPT to write code for you, and then using stackoverflow to find someone to debug your code, I suggest that you set aside 2 or 3 hours and carefully work your way through all of the [**mplfinance tutorials**](https://github.com/matplotlib/mplfinance#tutorials) <-click here. That will certainly answer most, if not all, of the questions you have. – Daniel Goldfarb May 31 '23 at 11:42
True. I admit that my knowledge of debugging is poor. Thank you for drawing my attention to this important area. I am now spending time to pick up debugging concepts. Regarding your observation of using ChatGPT to write code for me, I politely submit that it is not true. Kindly note that I am a retired Chemical Technologist of 68 yrs and my programming requirement is limited to my trading needs. I do not hope to become an expert professional python coder. To that extent, I will work hard and expect cooperation from peers like you. Regards. – DrSaaie K B Jun 03 '23 at 15:17

mplfinance: 'secondary_y=True' not working. Not plot multiple lines/candles due to very wide Y-Axis data ranges

1 Answers1