Merge pull request #1 from fatzh/master

Fixed dataframe generation with duplicated columns

pandas_ods_reader/parser.py

@@ -1,6 +1,7 @@
 """Imports an ods file into a DataFrame object"""
 import ezodf
 import pandas as pd
+from collections import OrderedDict
 
 from .tools import sanitize_df
 
@@ -15,22 +16,35 @@ def load_ods(doc, sheet_id, headers=True, columns=None):
             raise ValueError("There is no sheet named {}".format(sheet_id))
         sheet_id = sheets.index(sheet_id) + 1
     sheet = doc.sheets[sheet_id - 1]
-    df_dict = {}
+    df_dict = OrderedDict()
     col_index = {}
     for i, row in enumerate(sheet.rows()):
         # row is a list of cells
         if headers and i == 0 and not columns:
             # columns as lists in a dictionary
-            df_dict = {cell.value: [] for cell in row if cell.value}
+            columns = []
+            for cell in row:
+                if cell.value:
+                    if cell.value not in columns:
+                        columns.append(cell.value)
+                    else:
+                        # add count to column name
+                        idx = 1
+                        while "{}.{}".format(cell.value, idx) in columns:
+                            idx +=1
+                        columns.append("{}.{}".format(cell.value, idx))
+
+            df_dict = OrderedDict((column, []) for column in columns)
             # create index for the column headers
             col_index = {
-                j: cell.value for j, cell in enumerate(row) if cell.value}
+                j: column for j, column in enumerate(columns)
+            }
             continue
         elif i == 0:
             columns = columns if columns else (
                 [f"column_{j}" for j in range(len(row))])
             # columns as lists in a dictionary
-            df_dict = {column: [] for column in columns}
+            df_dict = OrderedDict((column, []) for column in columns)
             # create index for the column headers
             col_index = {j: column for j, column in enumerate(columns)}
             if headers:
@@ -41,7 +55,6 @@ def load_ods(doc, sheet_id, headers=True, columns=None):
                 df_dict[col_index[j]].append(cell.value)
             else:
                 continue
-    # and convert to a DataFrame
     df = pd.DataFrame(df_dict)
     return df
 

pandas_ods_reader/tests/test_read_ods.py

@@ -11,6 +11,7 @@ rsc = os.path.join(root, "rsc")
 
 header_file = "example_headers.ods"
 no_header_file = "example_no_headers.ods"
+duplicated_column_names_file = "example_duplicated_column_names.ods"
 col_len_file = "example_col_lengths.ods"
 
 
@@ -51,7 +52,13 @@ class TestOdsReader(object):
         df = read_ods(path, 1, headers=False, columns=columns)
         assert list(df.columns) == columns
         assert len(df) == 10
-        assert (len(df.columns) == 5)
+
+    def test_duplicated_column_names(self):
+        path = os.path.join(rsc, duplicated_column_names_file)
+        df = read_ods(path, 1)
+        assert isinstance(df, pd.DataFrame)
+        assert len(df.columns) == 4
+        assert "website.1" in df.columns
 
     def test_header_file_col_len(self):
         path = os.path.join(rsc, col_len_file)