diff --git a/saqc/core/translation/dmpscheme.py b/saqc/core/translation/dmpscheme.py
index bf35b7b895c04cda129f698763ebaf9f711ae3ab..6b9a9a5d383069f67c2f178df0e04337be86d286 100644
--- a/saqc/core/translation/dmpscheme.py
+++ b/saqc/core/translation/dmpscheme.py
@@ -17,6 +17,7 @@ import pandas as pd
from saqc import BAD, DOUBTFUL, GOOD, UNFLAGGED
from saqc.core import Flags, History
from saqc.core.translation.basescheme import BackwardMap, ForwardMap, MappingScheme
+from saqc.lib.tools import getUnionIndex
_QUALITY_CAUSES = [
"",
@@ -133,7 +134,7 @@ class DmpScheme(MappingScheme):
tflags = super().toExternal(flags, attrs=attrs)
out = pd.DataFrame(
- index=reduce(lambda x, y: x.union(y), tflags.indexes).sort_values(),
+ index=getUnionIndex(tflags),
columns=pd.MultiIndex.from_product([flags.columns, _QUALITY_LABELS]),
)
diff --git a/saqc/core/translation/positionalscheme.py b/saqc/core/translation/positionalscheme.py
index f503b0fe379019d62c40c9ce9d76bbaa21545f09..fffc988341204fd21d2136d7db5f3a1990de0bf2 100644
--- a/saqc/core/translation/positionalscheme.py
+++ b/saqc/core/translation/positionalscheme.py
@@ -91,7 +91,7 @@ class PositionalScheme(MappingScheme):
for field in flags.columns:
thist = flags.history[field].hist.replace(self._BACKWARD).astype(float)
# concatenate the single flag values
- ncols = thist.shape[-1]
+ ncols = len(thist.columns)
init = 9 * 10**ncols
bases = 10 ** np.arange(ncols - 1, -1, -1)
diff --git a/saqc/funcs/curvefit.py b/saqc/funcs/curvefit.py
index c0a86fd5185869c181af20b45b7259b1cc28801a..865550c8bad4d3eaedd5b4fda221544ec21e492e 100644
--- a/saqc/funcs/curvefit.py
+++ b/saqc/funcs/curvefit.py
@@ -212,7 +212,7 @@ def _fitPolynomial(
window = int(window - 1)
if min_periods is None:
min_periods = window
- if to_fit.shape[0] < 200000:
+ if len(to_fit) < 200000:
numba = False
else:
numba = True
diff --git a/saqc/funcs/drift.py b/saqc/funcs/drift.py
index d3555ac51ab8d7989aec5b34440a40e18677432b..2545b7d125d2525930a05956dfc6bf0f3b4819a2 100644
--- a/saqc/funcs/drift.py
+++ b/saqc/funcs/drift.py
@@ -138,12 +138,12 @@ class DriftMixin:
"""
fields = toSequence(field)
- data_to_flag = self._data[fields].to_df()
- data_to_flag.dropna(inplace=True)
+ data = self._data[fields].to_df()
+ data.dropna(inplace=True)
- segments = data_to_flag.groupby(pd.Grouper(freq=freq))
+ segments = data.groupby(pd.Grouper(freq=freq))
for segment in segments:
- if segment[1].shape[0] <= 1:
+ if len(segment[1]) <= 1:
continue
drifters = detectDeviants(
@@ -223,11 +223,11 @@ class DriftMixin:
if reference not in fields:
fields.append(reference)
- data_to_flag = self._data[fields].to_df().dropna()
+ data = self._data[fields].to_df().dropna()
- segments = data_to_flag.groupby(pd.Grouper(freq=freq))
+ segments = data.groupby(pd.Grouper(freq=freq))
for segment in segments:
- if segment[1].shape[0] <= 1:
+ if len(segment[1]) <= 1:
continue
for i in range(len(fields)):
@@ -346,7 +346,7 @@ class DriftMixin:
drift_frame = pd.DataFrame(d, index=to_correct_clean.index)
# group the drift frame
- for k in range(0, maint_data.shape[0] - 1):
+ for k in range(0, len(maint_data) - 1):
# assign group numbers for the timespans in between one maintenance ending and the beginning of the next
# maintenance time itself remains np.nan assigned
drift_frame.loc[
@@ -460,13 +460,13 @@ class DriftMixin:
first_valid = np.array(
[
~pd.isna(para_dict[unique_successive[i]]).any()
- for i in range(0, unique_successive.shape[0])
+ for i in range(0, len(unique_successive))
]
)
first_valid = np.where(first_normal & first_valid)[0][0]
last_valid = 1
- for k in range(0, unique_successive.shape[0]):
+ for k in range(0, len(unique_successive)):
if unique_successive[k] < 0 & (
not pd.isna(para_dict[unique_successive[k]]).any()
):
@@ -757,8 +757,8 @@ def _assignRegimeAnomaly(
plateaus = detectDeviants(cluster_dios, metric, spread, frac, method, "samples")
if set_flags:
- for p in plateaus:
- flags[cluster_dios.iloc[:, p].index, field] = flag
+ for p, cols in zip(plateaus, cluster_dios.columns[plateaus]):
+ flags[cluster_dios[cols].index, field] = flag
if set_cluster:
for p in plateaus:
diff --git a/saqc/funcs/generic.py b/saqc/funcs/generic.py
index e70b4c17e4ff9a41f5170acff21f1ed05be14e3a..803edfdaf8c481841db08731b854c0957f075381 100644
--- a/saqc/funcs/generic.py
+++ b/saqc/funcs/generic.py
@@ -15,7 +15,7 @@ import pandas as pd
from saqc import BAD, FILTER_ALL
from saqc.core import DictOfSeries, Flags, History, register
from saqc.core.register import _maskData
-from saqc.lib.tools import isflagged, toSequence
+from saqc.lib.tools import isAllBoolean, isflagged, toSequence
from saqc.lib.types import GenericFunction, PandasLike
from saqc.parsing.environ import ENVIRONMENT
@@ -51,7 +51,7 @@ def _prepare(
for f in fchunk.columns:
fchunk.history[f] = flags.history[f]
dchunk, _ = _maskData(
- data=data.loc[:, columns].copy(), flags=fchunk, columns=columns, thresh=dfilter
+ data=data[columns].copy(), flags=fchunk, columns=columns, thresh=dfilter
)
return dchunk, fchunk.copy()
@@ -169,7 +169,7 @@ class GenericMixin:
# update data & flags
for i, col in enumerate(targets):
- datacol = result.iloc[:, i]
+ datacol = result[result.columns[i]]
self._data[col] = datacol
if col not in self._flags:
@@ -272,7 +272,7 @@ class GenericMixin:
f"the generic function returned {len(result.columns)} field(s), but only {len(targets)} target(s) were given"
)
- if not result.empty and not (result.dtypes == bool).all():
+ if not result.empty and not isAllBoolean(result):
raise TypeError(f"generic expression does not return a boolean array")
meta = {
@@ -288,7 +288,7 @@ class GenericMixin:
# update flags & data
for i, col in enumerate(targets):
- maskcol = result.iloc[:, i]
+ maskcol = result[result.columns[i]]
# make sure the column exists
if col not in self._flags:
diff --git a/saqc/funcs/outliers.py b/saqc/funcs/outliers.py
index 182205ec39429b316fb9dba93ef7acfc7f1c25fd..03c50820f89542d883053d8a9894e1224592d8ad 100644
--- a/saqc/funcs/outliers.py
+++ b/saqc/funcs/outliers.py
@@ -131,14 +131,14 @@ class OutliersMixin:
return self
if not window:
- window = scores.shape[0]
+ window = len(scores)
if isinstance(window, str):
partitions = scores.groupby(pd.Grouper(freq=window))
else:
grouper_series = pd.Series(
- data=np.arange(0, scores.shape[0]), index=scores.index
+ data=np.arange(0, len(scores)), index=scores.index
)
grouper_series = grouper_series.transform(
lambda x: int(np.floor(x / window))
@@ -147,10 +147,10 @@ class OutliersMixin:
# calculate flags for every partition
for _, partition in partitions:
- if partition.empty | (partition.shape[0] < min_periods):
+ if partition.empty | (len(partition) < min_periods):
continue
- sample_size = partition.shape[0]
+ sample_size = len(partition)
sorted_i = partition.values.argsort()
resids = partition.values[sorted_i]
@@ -931,7 +931,7 @@ class OutliersMixin:
# period number defined test intervals
if isinstance(window, int):
grouper_series = pd.Series(
- data=np.arange(0, datcol.shape[0]), index=datcol.index
+ data=np.arange(0, len(datcol)), index=datcol.index
)
grouper_series_lagged = grouper_series + (window / 2)
grouper_series = grouper_series.transform(lambda x: x // window)
@@ -1036,7 +1036,7 @@ class OutliersMixin:
fields = toSequence(field)
- df = self._data[fields].loc[self._data[fields].index_of("shared")].to_df()
+ df = self._data[fields].to_df(how="inner")
if isinstance(method, str):
if method == "modZscore":
@@ -1265,7 +1265,7 @@ def _evalStrayLabels(
if reduction_drop_flagged:
test_slice = test_slice.drop(to_flag_frame.index, errors="ignore")
- if test_slice.shape[0] < reduction_min_periods:
+ if len(test_slice) < reduction_min_periods:
to_flag_frame.loc[index[1], var] = True
continue
diff --git a/saqc/funcs/resampling.py b/saqc/funcs/resampling.py
index 4d2a59e96a06ac8b0eba3e9cf8898436c8f47412..ebc02bd4bf0ad1e890050a42138c8c36fcefa253 100644
--- a/saqc/funcs/resampling.py
+++ b/saqc/funcs/resampling.py
@@ -270,13 +270,9 @@ class ResamplingMixin:
datcol = self._data[field]
- # workaround for #GL-333
if datcol.empty:
- if self._data.itype is None:
- index = pd.DatetimeIndex([])
- else:
- index = self._data.itype.min_pdindex
- datcol = pd.Series(index=index, dtype=datcol.dtype)
+ # see for #GL-374
+ datcol = pd.Series(index=pd.DatetimeIndex([]), dtype=datcol.dtype)
freq = evalFreqStr(freq, freq_check, datcol.index)
diff --git a/saqc/funcs/scores.py b/saqc/funcs/scores.py
index 1846cab2b40f8b4837acce6b0ff11c581d2691f3..9e3d0bc4ba3f3a6615681f0840e4e3eb25031f4c 100644
--- a/saqc/funcs/scores.py
+++ b/saqc/funcs/scores.py
@@ -180,11 +180,8 @@ class ScoresMixin:
target = target[0]
fields = toSequence(field)
- val_frame = self._data[fields].copy()
- score_index = val_frame.index_of("shared")
- score_ser = pd.Series(np.nan, index=score_index, name=target)
-
- val_frame = val_frame.loc[val_frame.index_of("shared")].to_df()
+ val_frame = self._data[fields].copy().to_df(how="inner")
+ score_ser = pd.Series(np.nan, index=val_frame.index, name=target)
val_frame.dropna(inplace=True)
if val_frame.empty:
@@ -192,23 +189,23 @@ class ScoresMixin:
# partitioning
if not freq:
- freq = val_frame.shape[0]
+ freq = len(val_frame.index)
if isinstance(freq, str):
grouper = pd.Grouper(freq=freq)
else:
grouper = pd.Series(
- data=np.arange(0, val_frame.shape[0]), index=val_frame.index
+ data=np.arange(0, len(val_frame)), index=val_frame.index
)
grouper = grouper.transform(lambda x: int(np.floor(x / freq)))
partitions = val_frame.groupby(grouper)
for _, partition in partitions:
- if partition.empty or (partition.shape[0] < min_periods):
+ if partition.empty or (len(partition) < min_periods):
continue
- sample_size = partition.shape[0]
+ sample_size = len(partition)
nn_neighbors = min(n, max(sample_size, 2) - 1)
dist, *_ = kNN(
partition.values, nn_neighbors, algorithm=method, metric=metric, p=p
diff --git a/saqc/funcs/tools.py b/saqc/funcs/tools.py
index 1036b7207c674f116ed24a34645b23478ead73ad..d8307a4597cbd2bd0eb2b90d0c59570180872d72 100644
--- a/saqc/funcs/tools.py
+++ b/saqc/funcs/tools.py
@@ -214,13 +214,14 @@ class ToolsMixin:
mask = periodicMask(datcol_idx, start, end, ~closed)
elif mode == "selection_field":
idx = self._data[selection_field].index.intersection(datcol_idx)
- mask = self._data.loc[idx, selection_field]
+ mask = self._data[selection_field].loc[idx]
else:
raise ValueError(
"Keyword passed as masking mode is unknown ({})!".format(mode)
)
- self._data.aloc[mask, field] = np.nan
+ mask = mask.reindex(self._data[field].index, fill_value=False).astype(bool)
+ self._data[field].loc[mask] = np.nan
self._flags[mask, field] = UNFLAGGED
return self
@@ -301,7 +302,7 @@ class ToolsMixin:
level = kwargs.get("flag", UNFLAGGED)
if dfilter < np.inf:
- data.loc[flags[field] >= dfilter, field] = np.nan
+ data[field].loc[flags[field] >= dfilter] = np.nan
if store_kwargs is None:
store_kwargs = {}
diff --git a/saqc/funcs/transformation.py b/saqc/funcs/transformation.py
index 0fe642131f21b2279bffa3ffc89b43dc3a060e3a..1e4121b8487eb3d5581b3001e342b146dfa23672 100644
--- a/saqc/funcs/transformation.py
+++ b/saqc/funcs/transformation.py
@@ -52,14 +52,12 @@ class TransformationMixin:
val_ser = self._data[field].copy()
# partitioning
if not freq:
- freq = val_ser.shape[0]
+ freq = len(val_ser)
if isinstance(freq, str):
grouper = pd.Grouper(freq=freq)
else:
- grouper = pd.Series(
- data=np.arange(0, val_ser.shape[0]), index=val_ser.index
- )
+ grouper = pd.Series(data=np.arange(0, len(val_ser)), index=val_ser.index)
grouper = grouper.transform(lambda x: int(np.floor(x / freq)))
partitions = val_ser.groupby(grouper)
diff --git a/saqc/lib/tools.py b/saqc/lib/tools.py
index 3a454bd181f6c938ee21c136c40cd25a7cdc7772..8bf96b839e1121b96ef300e153eee3e72e9d0b07 100644
--- a/saqc/lib/tools.py
+++ b/saqc/lib/tools.py
@@ -8,10 +8,11 @@
from __future__ import annotations
import collections
+import functools
import itertools
import re
import warnings
-from typing import Callable, Collection, List, Sequence, TypeVar, Union
+from typing import Any, Callable, Collection, List, Sequence, TypeVar, Union
import numpy as np
import pandas as pd
@@ -374,7 +375,9 @@ def detectDeviants(
dist_mat = np.zeros((var_num, var_num))
combs = list(itertools.combinations(range(0, var_num), 2))
for i, j in combs:
- dist = metric(data.iloc[:, i].values, data.iloc[:, j].values)
+ d_i = data[data.columns[i]]
+ d_j = data[data.columns[j]]
+ dist = metric(d_i.values, d_j.values)
dist_mat[i, j] = dist
condensed = np.abs(dist_mat[tuple(zip(*combs))])
@@ -386,7 +389,8 @@ def detectDeviants(
elif population == "samples":
counts = {cluster[j]: 0 for j in range(0, var_num)}
for c in range(var_num):
- counts[cluster[c]] += data.iloc[:, c].dropna().shape[0]
+ field = data.columns[c]
+ counts[cluster[c]] += data[field].dropna().shape[0]
pop_num = np.sum(list(counts.values()))
else:
raise ValueError(
@@ -557,3 +561,32 @@ def isflagged(flagscol: A, thresh: float) -> A:
return flagscol > UNFLAGGED
return flagscol >= thresh
+
+
+def getUnionIndex(obj, default: pd.DatetimeIndex | None = None):
+ assert hasattr(obj, "columns")
+ if default is None:
+ default = pd.DatetimeIndex([])
+ indexes = [obj[k].index for k in obj.columns]
+ if indexes:
+ return functools.reduce(pd.Index.union, indexes).sort_values()
+ return default
+
+
+def getSharedIndex(obj, default: pd.DatetimeIndex | None = None):
+ assert hasattr(obj, "columns")
+ if default is None:
+ default = pd.DatetimeIndex([])
+ indexes = [obj[k].index for k in obj.columns]
+ if indexes:
+ return functools.reduce(pd.Index.intersection, indexes).sort_values()
+ return default
+
+
+def isAllBoolean(obj: Any):
+ if not hasattr(obj, "columns"):
+ return pd.api.types.is_bool_dtype(obj)
+ for c in obj.columns:
+ if not pd.api.types.is_bool_dtype(obj[c]):
+ return False
+ return True
diff --git a/tests/common.py b/tests/common.py
index c82b5d2bf791a499dd9f5345ae9b3f69d0f18580..819f5d4f38cc96d1d025da5ec5654c0fc46a9258 100644
--- a/tests/common.py
+++ b/tests/common.py
@@ -22,14 +22,14 @@ def initData(
if rows is None:
freq = freq or "1h"
- di = DictOfSeries(itype="datetime")
+ di = dict()
dates = pd.date_range(start=start_date, end=end_date, freq=freq, periods=rows)
dummy = np.arange(len(dates))
for col in range(1, cols + 1):
di[f"var{col}"] = pd.Series(data=dummy * col, index=dates)
- return di
+ return DictOfSeries(di)
def dummyHistory(hist: pd.DataFrame = None, meta: list = None):
diff --git a/tests/funcs/test_constants_detection.py b/tests/funcs/test_constants_detection.py
index 29a5a251458cc225fa012652b2c509cfd39b866f..8cbf3290d58b2b76f80c8255aa874533b669a3cd 100644
--- a/tests/funcs/test_constants_detection.py
+++ b/tests/funcs/test_constants_detection.py
@@ -19,7 +19,8 @@ def data():
constants_data = initData(
1, start_date="2011-01-01 00:00:00", end_date="2011-01-01 03:00:00", freq="5min"
)
- constants_data.iloc[5:25] = 200
+ for c in constants_data.columns:
+ constants_data[c].iloc[5:25] = 200
return constants_data
diff --git a/tests/funcs/test_functions.py b/tests/funcs/test_functions.py
index 9ae18b532101c1822140aaf502913719644ddd33..991481575dd465368e82f5164956b3854b018538 100644
--- a/tests/funcs/test_functions.py
+++ b/tests/funcs/test_functions.py
@@ -53,8 +53,8 @@ def test_flagRange(data, field):
def test_flagSesonalRange(data, field):
- data.iloc[::2] = 0
- data.iloc[1::2] = 50
+ data[field].iloc[::2] = 0
+ data[field].iloc[1::2] = 50
nyears = len(data[field].index.year.unique())
tests = [
@@ -89,7 +89,7 @@ def test_flagSesonalRange(data, field):
start = f"{test['startmonth']:02}-{test['startday']:02}T00:00:00"
end = f"{test['endmonth']:02}-{test['endday']:02}T00:00:00"
- qc = qc.copyField(field, field + "_masked")
+ qc = qc.copyField(field, newfield)
qc = qc.selectTime(
newfield,
mode="periodic",
@@ -128,12 +128,12 @@ def test_forceFlags(data, field):
def test_flagIsolated(data, field):
flags = initFlagsLike(data)
- d_len = data.shape[0][0]
- data.iloc[1:3, 0] = np.nan
- data.iloc[4:5, 0] = np.nan
+ d_len = len(data[field].index)
+ data[field].iloc[1:3] = np.nan
+ data[field].iloc[4:5] = np.nan
flags[data[field].index[5:6], field] = BAD
- data.iloc[11:13, 0] = np.nan
- data.iloc[15:17, 0] = np.nan
+ data[field].iloc[11:13] = np.nan
+ data[field].iloc[15:17] = np.nan
# data flags
# 2016-01-01 0.0 -inf
diff --git a/tests/funcs/test_generic_api_functions.py b/tests/funcs/test_generic_api_functions.py
index e2181e9bd30d5c3c75472b8a06cd3b4cc062a321..18eb41240fb4a8aceb63ae99ac90af0548be38e9 100644
--- a/tests/funcs/test_generic_api_functions.py
+++ b/tests/funcs/test_generic_api_functions.py
@@ -106,20 +106,39 @@ def test_overwriteFieldFlagGeneric(data, fields, func):
@pytest.mark.parametrize(
- "targets, func",
+ "data, targets, func, expected_data",
[
- (["tmp"], lambda x, y: x + y),
- (["tmp1", "tmp2"], lambda x, y: (x + y, y * 2)),
+ (
+ DictOfSeries(dict(a=pd.Series([1.0, 2.0]), b=pd.Series([10.0, 20.0]))),
+ ["t1"],
+ lambda x, y: x + y,
+ DictOfSeries(
+ dict(
+ a=pd.Series([1.0, 2.0]),
+ b=pd.Series([10.0, 20.0]),
+ t1=pd.Series([11.0, 22.0]),
+ )
+ ),
+ ),
+ (
+ DictOfSeries(dict(a=pd.Series([1.0, 2.0]), b=pd.Series([10.0, 20.0]))),
+ ["t1", "t2"],
+ lambda x, y: (x + y, y * 2),
+ DictOfSeries(
+ dict(
+ a=pd.Series([1.0, 2.0]),
+ b=pd.Series([10.0, 20.0]),
+ t1=pd.Series([11.0, 22.0]),
+ t2=pd.Series([20.0, 40.0]),
+ )
+ ),
+ ),
],
)
-def test_writeTargetProcGeneric(data, targets, func):
- fields = ["var1", "var2"]
+def test_writeTargetProcGeneric(data, targets, func, expected_data):
+ fields = data.columns.tolist()
dfilter = 128
- expected_data = DictOfSeries(
- func(*[data[f] for f in fields]), columns=toSequence(targets)
- ).squeeze()
-
expected_meta = {
"func": "procGeneric",
"args": (fields, targets),
@@ -140,7 +159,7 @@ def test_writeTargetProcGeneric(data, targets, func):
dfilter=dfilter,
label="generic",
)
- assert (expected_data == res.data[targets].squeeze()).all(axis=None)
+ assert (expected_data == res.data).all(axis=None)
# check that new histories where created
for target in targets:
assert res._flags.history[target].hist.iloc[0].isna().all()
@@ -148,17 +167,24 @@ def test_writeTargetProcGeneric(data, targets, func):
@pytest.mark.parametrize(
- "fields, func",
+ "data, fields, func, expected_data",
[
- (["var1"], lambda x: x * 2),
- (["var1", "var2"], lambda x, y: (x + y, y * 2)),
+ (
+ DictOfSeries(dict(a=pd.Series([1.0, 2.0]), b=pd.Series([10.0, 20.0]))),
+ ["a"],
+ lambda x: x * 2,
+ DictOfSeries(dict(a=pd.Series([2.0, 4.0]), b=pd.Series([10.0, 20.0]))),
+ ),
+ (
+ DictOfSeries(dict(a=pd.Series([1.0, 2.0]), b=pd.Series([10.0, 20.0]))),
+ ["a", "b"],
+ lambda x, y: (x + y, y * 2),
+ DictOfSeries(dict(a=pd.Series([11.0, 22.0]), b=pd.Series([20.0, 40.0]))),
+ ),
],
)
-def test_overwriteFieldProcGeneric(data, fields, func):
+def test_overwriteFieldProcGeneric(data, fields, func, expected_data):
dfilter = 128
- expected_data = DictOfSeries(
- func(*[data[f] for f in fields]), columns=fields
- ).squeeze()
expected_meta = {
"func": "procGeneric",
@@ -176,7 +202,7 @@ def test_overwriteFieldProcGeneric(data, fields, func):
)
res = saqc.processGeneric(field=fields, func=func, dfilter=dfilter, label="generic")
- assert (expected_data == res.data[fields].squeeze()).all(axis=None)
+ assert (expected_data == res.data).all(axis=None)
# check that the histories got appended
for field in fields:
assert (res._flags.history[field].hist[0] == 127.0).all()
diff --git a/tests/funcs/test_generic_config_functions.py b/tests/funcs/test_generic_config_functions.py
index 0dc9d20ec672887a4977ad7c2cead67f1511473b..194f62d88de4e9ac9612eeff4aa0ed829153d62e 100644
--- a/tests/funcs/test_generic_config_functions.py
+++ b/tests/funcs/test_generic_config_functions.py
@@ -113,7 +113,7 @@ def test_arithmeticOperators(data):
def test_nonReduncingBuiltins(data):
var1, *_ = data.columns
- data = data.iloc[1:10, 0]
+ data = data[var1].iloc[1:10]
flags = Flags({var1: pd.Series(UNFLAGGED, index=data.index)})
tests = [
diff --git a/tests/fuzzy/test_masking.py b/tests/fuzzy/test_masking.py
index 472601037279556a5f069ff703ea4cb95a789c6e..ef7d54e968688ca7a026bcd4261855821b11bfe0 100644
--- a/tests/fuzzy/test_masking.py
+++ b/tests/fuzzy/test_masking.py
@@ -61,7 +61,7 @@ def test_flagsMutationPreventsUnmasking(data_field_flags):
data_masked, mask = _maskData(data_in, flags, columns=[field], thresh=UNFLAGGED)
flags[:, field] = UNFLAGGED
data_out = _unmaskData(data_masked, mask)
- assert (data_out.loc[flags[field] == BAD, field].isna()).all(axis=None)
+ assert (data_out[field].loc[flags[field] == BAD].isna()).all(axis=None)
@pytest.mark.slow