[PYTHON] Acquire multiple Excels at once with glob (Borace machine learning prediction result, confirmation for June)

Introduction

• The code explanation itself in this article is for beginners in python programming.

On the boat race triple forecast site "Today, do you have a good forecast?", the hit rate and recovery rate of daily race forecasts are disclosed without hiding. However, I wanted to put together the daily forecast results once a month, so I thought about processing like the title.

My situation

We organize the results of boat races and machine learning predictions and save them daily in a format such as "result_2020mmdd.csv". I want to put these files together once a month and visualize the results ...

And I want to omit the files like result_202006 .. ** _ test ** .csv which are mixed in some places as shown in the figure below because they are for testing.

The code looks like this.

It's very simple.

import pandas as pd
import glob

csv_files = glob.glob("predict/result/result_202006??.csv")
filelist = []
for file in csv_files:
    filelist.append(pd.read_csv(file))

df = pd.concat(filelist)

When I get the path name with glob, I use result_202006 ** ?? ** .csv. One? Will take charge of any one character.

csv_files has a nice set of files + pathnames.

Append one by one with a for statement, and finally make it into a DataFrame and complete! is.

By the way, how was the result in June ..

I checked the result using the data frame I created earlier. The triple-unit hit rate was 10% as designed, but the recovery rate is just over 80%. (However, looking at other free forecasts, it may seem like you can still fight relatively well at this point.)

Here is the result of organizing with pivot_table after summarizing. ** It is interesting that the boat races that are easy to hit and the boat races that do not hit at all are clear. ** ** Let's buy only the boat races that are easy to hit this month ..