fix: duplicate split keys

FelixAbrahamsson · FelixAbrahamsson · commit 526ec52361ca · 2021-05-03T17:41:47.000+02:00
diff --git a/datastream/tools/split_dataframes.py b/datastream/tools/split_dataframes.py
@@ -58,7 +58,7 @@ def split_dataframes(
             )
 
         last_split_name, _ = split_proportions[-1]
-        split[last_split_name] += unassigned(dataframe[key_column], split)
+        split[last_split_name] += unassigned(key_dataframe[key_column], split)
 
         if filepath is not None:
             filepath.parent.mkdir(parents=True, exist_ok=True)

Original file line number	Diff line number	Diff line change
`@@ -58,7 +58,7 @@ def split_dataframes(`
`58`	`58`	`)`
`59`	`59`
`60`	`60`	`last_split_name, _ = split_proportions[-1]`
`61`		`- split[last_split_name] += unassigned(dataframe[key_column], split)`
	`61`	`+ split[last_split_name] += unassigned(key_dataframe[key_column], split)`
`62`	`62`
`63`	`63`	`if filepath is not None:`
`64`	`64`	`filepath.parent.mkdir(parents=True, exist_ok=True)`