{"id":3764,"date":"2023-07-15T16:32:02","date_gmt":"2023-07-15T16:32:02","guid":{"rendered":"https:\/\/statorials.org\/pl\/pandy-groupby-diff\/"},"modified":"2023-07-15T16:32:02","modified_gmt":"2023-07-15T16:32:02","slug":"pandy-groupby-diff","status":"publish","type":"post","link":"https:\/\/statorials.org\/pl\/pandy-groupby-diff\/","title":{"rendered":"Pandy: jak u\u017cywa\u0107 groupby z diff"},"content":{"rendered":"

<\/p>\n

\n

Aby u\u017cywa\u0107 funkcji groupby()<\/strong> z funkcj\u0105 diff()<\/strong> w pandach, mo\u017cesz u\u017cy\u0107 nast\u0119puj\u0105cej podstawowej sk\u0142adni:<\/span><\/p>\n

 df = df. sort_values<\/span> (by=[' group_var1<\/span> ', ' group_var2<\/span> '])\n\ndf[' diff<\/span> '] = df. groupby<\/span> ([' group_var1<\/span> '])[' values_var<\/span> ']. diff<\/span> (). fillna<\/span> ( 0<\/span> )\n<\/strong><\/pre>\n
 Ten konkretny przyk\u0142ad sortuje wiersze ramki DataFrame wed\u0142ug dw\u00f3ch okre\u015blonych zmiennych, nast\u0119pnie grupuje je wed\u0142ug group_var1<\/strong> i oblicza r\u00f3\u017cnic\u0119 mi\u0119dzy wierszami w kolumnie warto\u015bci_zmienna<\/strong> .<\/span><\/p>\n
 Zauwa\u017c, \u017ce fillna(0)<\/strong> m\u00f3wi pandom, aby wstawia\u0142y zero za ka\u017cdym razem, gdy warto\u015b\u0107 zmiennej grupowej zmienia si\u0119 pomi\u0119dzy kolejnymi wierszami w ramce DataFrame.<\/span><\/p>\n
 Poni\u017cszy przyk\u0142ad pokazuje, jak zastosowa\u0107 t\u0119 sk\u0142adni\u0119 w praktyce.<\/span><\/p>\n
 Przyk\u0142ad: Jak u\u017cywa\u0107 groupby z r\u00f3\u017cnic\u0105 w Pandach<\/strong><\/span><\/h2>\n
 Za\u0142\u00f3\u017cmy, \u017ce mamy nast\u0119puj\u0105c\u0105 ramk\u0119 danych pand, kt\u00f3ra zawiera ca\u0142kowit\u0105 sprzeda\u017c dokonan\u0105 przez dwa r\u00f3\u017cne sklepy w r\u00f3\u017cnych datach:<\/span><\/p>\n
 import<\/span> pandas as<\/span> pd\n\n#createDataFrame<\/span>\ndf = pd. DataFrame<\/span> ({' store<\/span> ': ['A', 'A', 'A', 'A', 'B', 'B', 'B', 'B'],\n                   ' date<\/span> ': pd. to_datetime<\/span> (['2022-01-01', '2022-01-02',\n                                           '2022-01-03', '2022-01-04',\n                                           '2022-01-01', '2022-01-02',\n                                           '2022-01-03', '2022-01-04']),\n                   ' sales<\/span> ': [12, 15, 24, 24, 14, 19, 12, 38]})\n\n#view DataFrame\n<\/span>print<\/span> (df)\n\n  blind date sales\n0 A 2022-01-01 12\n1 A 2022-01-02 15\n2 A 2022-01-03 24\n3 A 2022-01-04 24\n4 B 2022-01-01 14\n5 B 2022-01-02 19\n6 B 2022-01-03 12\n7 B 2022-01-04 38<\/b><\/pre>\n
 Za\u0142\u00f3\u017cmy teraz, \u017ce chcemy utworzy\u0107 now\u0105 kolumn\u0119 o nazwie sales_diff<\/strong> zawieraj\u0105c\u0105 r\u00f3\u017cnic\u0119 w warto\u015bciach sprzeda\u017cy pomi\u0119dzy kolejnymi datami, pogrupowanymi wed\u0142ug sklep\u00f3w.<\/span><\/p>\n
 W tym celu mo\u017cemy u\u017cy\u0107 nast\u0119puj\u0105cej sk\u0142adni:<\/span><\/p>\n
 #sort DataFrame by store and date\n<\/span>df = df. sort_values<\/span> (by=[' store<\/span> ', ' date<\/span> '])\n\n#create new column that contains difference between sales grouped by store\n<\/span>df[' sales_diff<\/span> '] = df. groupby<\/span> ([' store<\/span> '])[' sales<\/span> ']. diff<\/span> (). fillna<\/span> ( 0<\/span> )\n\n#view update DataFrame\n<\/span>print<\/span> (df)\n\n  store date sales sales_diff\n0 A 2022-01-01 12 0.0\n1 A 2022-01-02 15 3.0\n2 A 2022-01-03 24 9.0\n3 A 2022-01-04 24 0.0\n4 B 2022-01-01 14 0.0\n5 B 2022-01-02 19 5.0\n6 B 2022-01-03 12 -7.0\n7 B 2022-01-04 38 26.0\n<\/span><\/span><\/strong><\/pre>\n
 Nowa kolumna sales_diff<\/strong> zawiera r\u00f3\u017cnic\u0119 warto\u015bci sprzeda\u017cy pomi\u0119dzy kolejnymi datami, pogrupowanymi wed\u0142ug sklep\u00f3w.<\/span><\/p>\n
 Na przyk\u0142ad mo\u017cemy zobaczy\u0107:<\/span><\/p>\n