rgwfuncs 0.0.12__py3-none-any.whl → 0.0.13__py3-none-any.whl
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- rgwfuncs/df_lib.py +11 -10
- {rgwfuncs-0.0.12.dist-info → rgwfuncs-0.0.13.dist-info}/METADATA +1 -1
- rgwfuncs-0.0.13.dist-info/RECORD +8 -0
- rgwfuncs-0.0.12.dist-info/RECORD +0 -8
- {rgwfuncs-0.0.12.dist-info → rgwfuncs-0.0.13.dist-info}/LICENSE +0 -0
- {rgwfuncs-0.0.12.dist-info → rgwfuncs-0.0.13.dist-info}/WHEEL +0 -0
- {rgwfuncs-0.0.12.dist-info → rgwfuncs-0.0.13.dist-info}/entry_points.txt +0 -0
- {rgwfuncs-0.0.12.dist-info → rgwfuncs-0.0.13.dist-info}/top_level.txt +0 -0
rgwfuncs/df_lib.py
CHANGED
@@ -61,6 +61,7 @@ def docs(method_type_filter: Optional[str] = None) -> None:
|
|
61
61
|
# Print the entire docstring for the matching function
|
62
62
|
print(f"\n{name}:\n{docstring}")
|
63
63
|
|
64
|
+
|
64
65
|
def numeric_clean(df: pd.DataFrame, column_names: str, column_type: str, irregular_value_treatment: str) -> pd.DataFrame:
|
65
66
|
"""
|
66
67
|
Cleans the numeric columns based on specified treatments.
|
@@ -1629,8 +1630,8 @@ def union_join(df1: pd.DataFrame, df2: pd.DataFrame) -> pd.DataFrame:
|
|
1629
1630
|
ValueError: If the DataFrames do not have the same columns.
|
1630
1631
|
"""
|
1631
1632
|
# Inspect initial columns
|
1632
|
-
print("Initial df1 columns:", df1.columns)
|
1633
|
-
print("Initial df2 columns:", df2.columns)
|
1633
|
+
# print("Initial df1 columns:", df1.columns)
|
1634
|
+
# print("Initial df2 columns:", df2.columns)
|
1634
1635
|
|
1635
1636
|
# Standardize columns by adding missing columns filled with NaN
|
1636
1637
|
for col in df2.columns:
|
@@ -1641,8 +1642,8 @@ def union_join(df1: pd.DataFrame, df2: pd.DataFrame) -> pd.DataFrame:
|
|
1641
1642
|
if col not in df2:
|
1642
1643
|
df2[col] = pd.NA
|
1643
1644
|
|
1644
|
-
print("Standardized df1 columns:", df1.columns)
|
1645
|
-
print("Standardized df2 columns:", df2.columns)
|
1645
|
+
# print("Standardized df1 columns:", df1.columns)
|
1646
|
+
# print("Standardized df2 columns:", df2.columns)
|
1646
1647
|
|
1647
1648
|
# Check if columns match now
|
1648
1649
|
if set(df1.columns) != set(df2.columns):
|
@@ -1652,6 +1653,7 @@ def union_join(df1: pd.DataFrame, df2: pd.DataFrame) -> pd.DataFrame:
|
|
1652
1653
|
result_df = pd.concat([df1, df2], ignore_index=True).drop_duplicates()
|
1653
1654
|
return result_df
|
1654
1655
|
|
1656
|
+
|
1655
1657
|
def bag_union_join(df1: pd.DataFrame, df2: pd.DataFrame) -> pd.DataFrame:
|
1656
1658
|
"""
|
1657
1659
|
Perform a bag union join, concatenating the two DataFrames without dropping duplicates.
|
@@ -1667,8 +1669,8 @@ def bag_union_join(df1: pd.DataFrame, df2: pd.DataFrame) -> pd.DataFrame:
|
|
1667
1669
|
ValueError: If the DataFrames do not have the same columns.
|
1668
1670
|
"""
|
1669
1671
|
# Inspect initial columns
|
1670
|
-
print("Initial df1 columns:", df1.columns)
|
1671
|
-
print("Initial df2 columns:", df2.columns)
|
1672
|
+
# print("Initial df1 columns:", df1.columns)
|
1673
|
+
# print("Initial df2 columns:", df2.columns)
|
1672
1674
|
|
1673
1675
|
# Standardize columns by adding missing columns filled with NaN
|
1674
1676
|
for col in df2.columns:
|
@@ -1679,8 +1681,8 @@ def bag_union_join(df1: pd.DataFrame, df2: pd.DataFrame) -> pd.DataFrame:
|
|
1679
1681
|
if col not in df2:
|
1680
1682
|
df2[col] = pd.NA
|
1681
1683
|
|
1682
|
-
print("Standardized df1 columns:", df1.columns)
|
1683
|
-
print("Standardized df2 columns:", df2.columns)
|
1684
|
+
# print("Standardized df1 columns:", df1.columns)
|
1685
|
+
# print("Standardized df2 columns:", df2.columns)
|
1684
1686
|
|
1685
1687
|
# Ensure they have the same columns after standardizing
|
1686
1688
|
if set(df1.columns) != set(df2.columns):
|
@@ -1722,6 +1724,7 @@ def right_join(df1: pd.DataFrame, df2: pd.DataFrame, left_on: str, right_on: str
|
|
1722
1724
|
"""
|
1723
1725
|
return df1.merge(df2, how='right', left_on=left_on, right_on=right_on)
|
1724
1726
|
|
1727
|
+
|
1725
1728
|
def sync_dataframe_to_sqlite_database(db_path: str, tablename: str, df: pd.DataFrame) -> None:
|
1726
1729
|
"""
|
1727
1730
|
Processes and saves a DataFrame to an SQLite database, adding a timestamp column
|
@@ -1776,5 +1779,3 @@ def sync_dataframe_to_sqlite_database(db_path: str, tablename: str, df: pd.DataF
|
|
1776
1779
|
conn.execute(f"DROP TABLE IF EXISTS {tablename}")
|
1777
1780
|
# Rename the new table to the old table name
|
1778
1781
|
conn.execute(f"ALTER TABLE {new_table_name} RENAME TO {tablename}")
|
1779
|
-
|
1780
|
-
|
@@ -0,0 +1,8 @@
|
|
1
|
+
rgwfuncs/__init__.py,sha256=BP8Nh8ivyCCz8Ga-21JW3NWInJFOElKoIfRuioJRWbA,1076
|
2
|
+
rgwfuncs/df_lib.py,sha256=2JU1Z0wjBrpcmLuJgY6-Bi1OFXHyNk1XeRGESkDI934,64077
|
3
|
+
rgwfuncs-0.0.13.dist-info/LICENSE,sha256=7EI8xVBu6h_7_JlVw-yPhhOZlpY9hP8wal7kHtqKT_E,1074
|
4
|
+
rgwfuncs-0.0.13.dist-info/METADATA,sha256=TZoSn8ANcHG2IOAW1SwNbu3ZnPKxdWbKzoCAXLTB06w,32059
|
5
|
+
rgwfuncs-0.0.13.dist-info/WHEEL,sha256=In9FTNxeP60KnTkGw7wk6mJPYd_dQSjEZmXdBdMCI-8,91
|
6
|
+
rgwfuncs-0.0.13.dist-info/entry_points.txt,sha256=j-c5IOPIQ0252EaOV6j6STio56sbXl2C4ym_fQ0lXx0,43
|
7
|
+
rgwfuncs-0.0.13.dist-info/top_level.txt,sha256=aGuVIzWsKiV1f2gCb6mynx0zx5ma0B1EwPGFKVEMTi4,9
|
8
|
+
rgwfuncs-0.0.13.dist-info/RECORD,,
|
rgwfuncs-0.0.12.dist-info/RECORD
DELETED
@@ -1,8 +0,0 @@
|
|
1
|
-
rgwfuncs/__init__.py,sha256=BP8Nh8ivyCCz8Ga-21JW3NWInJFOElKoIfRuioJRWbA,1076
|
2
|
-
rgwfuncs/df_lib.py,sha256=mIOEvnqpkkHW2Az23_09Xt61nltTtG6uNW0b81aW5Mg,64060
|
3
|
-
rgwfuncs-0.0.12.dist-info/LICENSE,sha256=7EI8xVBu6h_7_JlVw-yPhhOZlpY9hP8wal7kHtqKT_E,1074
|
4
|
-
rgwfuncs-0.0.12.dist-info/METADATA,sha256=rgeOUbLIXmgb-JuqmAcdF2G2vBS7iWCdsop74rfuTfc,32059
|
5
|
-
rgwfuncs-0.0.12.dist-info/WHEEL,sha256=In9FTNxeP60KnTkGw7wk6mJPYd_dQSjEZmXdBdMCI-8,91
|
6
|
-
rgwfuncs-0.0.12.dist-info/entry_points.txt,sha256=j-c5IOPIQ0252EaOV6j6STio56sbXl2C4ym_fQ0lXx0,43
|
7
|
-
rgwfuncs-0.0.12.dist-info/top_level.txt,sha256=aGuVIzWsKiV1f2gCb6mynx0zx5ma0B1EwPGFKVEMTi4,9
|
8
|
-
rgwfuncs-0.0.12.dist-info/RECORD,,
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|