組合兩個for循環以提高效率

Question

我通過循環csv文件來制作一堆圖，然后基於groupby制作多個圖。 代碼如下：

import pandas as pd   
import matplotlib.pyplot as plt
from matplotlib.backends.backend_pdf import PdfPages


frame=pd.read_csv('C:\\')

pdf_files = {}
for group_name, group in frame.groupby(['Allotment','Year','Month','Day']):
    allotment,year,month,day = group_name
    if month not in pdf_files:
        pdf_files[allotment,month] = PdfPages(r'F:\Sheyenne\Statistics\IDL_stats\Allotment_histos\Month\SWIR32' + '_' + allotment + '_'+ month + '.pdf') 
    plot=group.plot(x='Percent', y='SWIR32', title=str(group_name)).get_figure()
    pdf_files[allotment,month].savefig(plot)
    plt.close(plot)

for key in pdf_files:
    pdf_files[key].close()

print "Done"  

但這返回一個錯誤，指出打開的文件太多。 我認為如果可以將兩個for循環組合為一個，則可以解決此問題，但是我不確定如何做到這一點。

Answer 1

您首先無法按['allotment', 'month']分組的任何原因['allotment', 'month']然后每個循環將只是一個pdf文件（最好with PdfPages(...) as pdf_file: ）

basename = r'F:\Sheyenne\Statistics\IDL_stats\Allotment_histos\Month\SWIR32'
for file_name, file in frame.groupby(['Allotment','Month']):
    allotment, month = file_name
    with PdfPages('{}_{}_{}.pdf'.format(basename, allotment, month)) as pdf_file:
        for group_name, group in file.groupby(['Allotment','Month','Year', 'Day']):
            plot = group.plot(x='Percent', y='SWIR32', title=str(group_name)).get_figure()
            pdf_file.savefig(plot)
            plt.close(plot)

Answer 2

這行得通嗎？

import pandas as pd   
import matplotlib.pyplot as plt
from matplotlib.backends.backend_pdf import PdfPages

frame=pd.read_csv('C:\\')

pdf_files = {}
for group_name, group in frame.groupby(['Allotment','Year','Month','Day']):
    allotment,year,month,day = group_name
    if month not in pdf_files:
        pdf_files[allotment,month] = PdfPages(r'F:\Sheyenne\Statistics\IDL_stats\Allotment_histos\Month\SWIR32' + '_' + allotment + '_'+ month + '.pdf') 
    plot=group.plot(x='Percent', y='SWIR32', title=str(group_name)).get_figure()
    pdf_files[allotment,month].savefig(plot)
    pdf_files[allotment,month].close()
    plt.close(plot)

print "Done"  

基本上，只需確保在完成編輯后關閉文件即可。