Add new check: matches_regex

bulwark/checks.py

@@ -408,6 +408,35 @@ def is_same_as(df, df_to_compare, **kwargs):
     return df
 
 
+def matches_regex(df, pattern, columns=None, **kwargs):
+    """Asserts that the values in `df`'s `columns` match `pattern`.
+
+    Args:
+        df (pd.DataFrame): Any pd.DataFrame.
+        pattern (str): Pattern to match against.
+        columns (list): A subset of columns to check for matching `pattern`.
+        **kwargs (dict): Keyword arguments passed through to pandas' ``Series.str.match``.
+
+    Returns:
+        Original `df`.
+
+    """
+    columns = columns if columns is not None else df.columns
+
+    non_matches_df = pd.DataFrame()
+
+    for col in columns:
+        matches = df[col].str.match(pattern, **kwargs).fillna(False)
+        if not matches.all():
+            non_matches_df[col] = ~matches
+
+    if not non_matches_df.empty:
+        msg = bad_locations(non_matches_df)
+        raise AssertionError(msg)
+
+    return df
+
+
 def multi_check(df, checks, warn=False):
     """Asserts that all checks pass.
 

tests/test_checks.py

@@ -1,4 +1,6 @@
 # -*- coding: utf-8 -*-
+import re
+
 import pytest
 import numpy as np
 import pandas as pd
@@ -395,6 +397,67 @@ def test_is_same_as_with_kwargs():
     tm.assert_frame_equal(df, result)
 
 
+def test_matches_regex():
+    df = pd.DataFrame({'A': ['aa', 'ab', 'ac'], 'B': ['ad', 'ae', 'a1']})
+    pattern = r'a.'
+    result = ck.matches_regex(df, pattern)
+    tm.assert_frame_equal(df, result)
+
+    result = dc.MatchesRegex(pattern)(_noop)(df)
+    tm.assert_frame_equal(df, result)
+
+    pattern = r'aa'
+    with pytest.raises(AssertionError):
+        ck.matches_regex(df, pattern)
+        dc.MatchesRegex(pattern)(_noop)(df)
+
+
+def test_matches_regex_with_columns():
+    df = pd.DataFrame({'A': ['aa', 'ab', 'ac'], 'B': ['ad', 'ae', 'a1']})
+    pattern = r'a.'
+    result = ck.matches_regex(df, pattern, columns=['A'])
+    tm.assert_frame_equal(df, result)
+
+    result = dc.MatchesRegex(pattern, columns=['A'])(_noop)(df)
+    tm.assert_frame_equal(df, result)
+
+    pattern = r'a[a-z]'
+    with pytest.raises(AssertionError):
+        ck.matches_regex(df, pattern, columns=['B'])
+        dc.MatchesRegex(pattern, columns=['B'])(_noop)(df)
+
+
+def test_matches_regex_with_kwargs():
+    df = pd.DataFrame({'A': ['aa', 'ab', 'ac'], 'B': ['ad', 'ae', 'a1']})
+    pattern = 'A.'
+    result = ck.matches_regex(df, pattern, case=False)
+    tm.assert_frame_equal(df, result)
+
+    result = dc.MatchesRegex(pattern, case=False)(_noop)(df)
+    tm.assert_frame_equal(df, result)
+
+    pattern = 'A.'
+    result = ck.matches_regex(df, pattern, flags=re.IGNORECASE)
+    tm.assert_frame_equal(df, result)
+
+    result = dc.MatchesRegex(pattern, flags=re.IGNORECASE)(_noop)(df)
+    tm.assert_frame_equal(df, result)
+
+    df = pd.DataFrame({'A': ['aa', 'ab', np.nan]})
+    pattern = 'a.'
+
+    # should raise because of the nan
+    with pytest.raises(AssertionError):
+        ck.matches_regex(df, pattern)
+        dc.MatchesRegex(pattern)(_noop)(df)
+
+    result = ck.matches_regex(df, pattern, na='az')
+    tm.assert_frame_equal(df, result)
+
+    result = dc.MatchesRegex(pattern, na='az')(_noop)(df)
+    tm.assert_frame_equal(df, result)
+
+
 def test_multi_check():
     df = pd.DataFrame({"a": [1, 2, 3], "b": [4, 5, 6]})
     result = ck.multi_check(df,