Trim input values. (#226)

datacommonsorg · Oct 10, 2023 · d383700 · d383700
1 parent a34d096
commit d383700
Show file tree

Hide file tree

Showing 2 changed files with 7 additions and 2 deletions.
diff --git a/simple/stats/importer.py b/simple/stats/importer.py
@@ -50,6 +50,7 @@ def __init__(
     def do_import(self) -> None:
         self._read_csv()
         self._drop_ignored_columns()
+        self._trim_values()
         self._resolve_entities()
         self._rename_columns()
 
@@ -63,6 +64,10 @@ def _drop_ignored_columns(self):
         if self.ignore_columns:
             self.df.drop(columns=self.ignore_columns, axis=1, inplace=True)
 
+    def _trim_values(self):
+        self.df = self.df.map(lambda value: value.strip()
+                              if isinstance(value, str) else value)
+
     def _rename_columns(self) -> None:
         df = self.df
         df.columns.values[0] = constants.COLUMN_DCID

diff --git a/simple/stats/sample/input/geoids.csv b/simple/stats/sample/input/geoids.csv
@@ -1,3 +1,3 @@
 geoId,year,ignore1,var1,var2,ignore2
-01,2021,foo,555,666,bar
-122,2022,#N/A,321,123,baz
+01,2021,foo,555, 666 ,bar
+122,2022,#N/A,321 , 123,baz