回答編集履歴

2023/04/11 03:36

投稿

スコア21527

answer CHANGED Viewed

@@ -5,29 +5,29 @@
 import json
 csv_data = '''
-階層1,階層2,階層3,所有者
+階層1,階層2,階層3,階層4,所有者
-果物,赤,りんご,A
+果物,赤,りんご,,A
-果物,赤,いちご,A
+果物,赤,いちご,甘い,A
-果物,黄色,ばなな,B
+果物,黄色,ばなな,,B
-果物,黄色,れもん,C
+果物,黄色,れもん,酸っぱい,C
-野菜,赤,にんじん,C
+野菜,赤,にんじん,,C
-野菜,赤,とまと,A
+野菜,赤,とまと,甘い,A
-野菜,緑,きゅうり,B
+野菜,緑,きゅうり,,B
-野菜,緑,せろり,A
+野菜,緑,せろり,,A
-その他,赤,,B
+その他,赤,,,B
-その他,黄色,,A
+その他,黄色,,,A
 '''
-df = pd.read_csv(io.StringIO(csv_data))
 #
 def nested_dict(df):
     l = len(df.columns)
     if l == 1:
-        return df.values[0][0]
+        return df.values[0][0] #if df.values.size == 1 else df.values.squeeze()
     if df.values[0][0] is np.nan:
+        idx = df.iloc[0].first_valid_index()
-        if l == 2:
+        if idx == df.columns[-1]:
-            return df.values[0][1]
+            return df.values[0][-1]
-        df = df.iloc[:,1:]
+        df = df.loc[:,idx:]
     grp = df.groupby(df.columns[0], sort=False, dropna=False)
     return {k: nested_dict(g.iloc[:,1:]) for k, g in grp}
@@ -39,17 +39,23 @@
 #   "果物": {
 #     "赤": {
 #       "りんご": "A",
+#       "いちご": {
-#       "いちご": "A"
+#         "甘い": "A"
+#       }
 #     },
 #     "黄色": {
 #       "ばなな": "B",
+#       "れもん": {
-#       "れもん": "C"
+#         "酸っぱい": "C"
+#       }
 #     }
 #   },
 #   "野菜": {
 #     "赤": {
 #       "にんじん": "C",
+#       "とまと": {
-#       "とまと": "A"
+#         "甘い": "A"
+#       }
 #     },
 #     "緑": {
 #       "きゅうり": "B",
@@ -61,4 +67,4 @@
 #     "黄色": "A"
 #   }
 # }
-```
+```

2023/04/11 01:24

投稿

melian

スコア21527

answer CHANGED Viewed

@@ -27,8 +27,7 @@
     if df.values[0][0] is np.nan:
         if l == 2:
             return df.values[0][1]
-        if l > 2:
-            df = df.iloc[:,1:]
+        df = df.iloc[:,1:]
     grp = df.groupby(df.columns[0], sort=False, dropna=False)
     return {k: nested_dict(g.iloc[:,1:]) for k, g in grp}

2023/04/10 12:32

投稿

melian

スコア21527

answer CHANGED Viewed

@@ -1,7 +1,7 @@
-※ 「その他」の部分が「期待する結果」とは異なります
 ```python
 import io
 import pandas as pd
+import numpy as np
 import json
 csv_data = '''
@@ -21,12 +21,18 @@
 #
 def nested_dict(df):
-    if len(df.columns) == 1:
+    l = len(df.columns)
+    if l == 1:
-        return df.values[0][0]
+        return df.values[0][0]
+    if df.values[0][0] is np.nan:
+        if l == 2:
+            return df.values[0][1]
+        if l > 2:
+            df = df.iloc[:,1:]
     grp = df.groupby(df.columns[0], sort=False, dropna=False)
     return {k: nested_dict(g.iloc[:,1:]) for k, g in grp}
-dic = nested_dict(df.ffill(axis=1))
+dic = nested_dict(df)
 js = json.dumps(dic, ensure_ascii=False, indent=2)
 print(js)
@@ -52,12 +58,8 @@
 #     }
 #   },
 #   "その他": {
-#     "赤": {
-#       "赤": "B"
+#     "赤": "B",
-#     },
-#     "黄色": {
-#       "黄色": "A"
+#     "黄色": "A"
-#     }
 #   }
 # }
-```
+```

2023/04/10 11:45

投稿

melian

スコア21527

answer CHANGED Viewed

@@ -1,11 +1,8 @@
 ※ 「その他」の部分が「期待する結果」とは異なります
 ```python
 import io
 import pandas as pd
-import numpy as np
 import json
-from functools import reduce
 csv_data = '''
 階層1,階層2,階層3,所有者
@@ -23,14 +20,14 @@
 df = pd.read_csv(io.StringIO(csv_data))
 #
-dic = df.assign(l3 = df[['階層2', '階層3', '所有者']]\
+def nested_dict(df):
-        .apply(lambda x: {(x[0] if x[1] is np.nan else x[1]): x[2]}, axis=1))\
+    if len(df.columns) == 1:
-        .pivot_table(index=['階層1'], columns=['階層2'], values=['l3'],
+        return df.values[0][0]
-                     aggfunc=lambda x: reduce(lambda a, b: dict(a, **b), x), sort=False)\
+    grp = df.groupby(df.columns[0], sort=False, dropna=False)
-        .droplevel(0, axis=1).to_dict(orient='index')
+    return {k: nested_dict(g.iloc[:,1:]) for k, g in grp}
-js = json.dumps({k1: {k2: v2 for k2, v2 in v1.items() if v2 is not np.nan} for k1, v1 in dic.items()},
+dic = nested_dict(df.ffill(axis=1))
-                ensure_ascii=False, indent=2)
+js = json.dumps(dic, ensure_ascii=False, indent=2)
 print(js)
 # {