df = df[df.duration_mean > 1]
df.drop('duration_mean', axis=1, inplace=True)
minus_day_cases = df.val.sum()
start_icds = df.dx_1.unique()
diagnosis_feats = df.columns[df.columns.str.startswith('dx_')]
for feat in diagnosis_feats:
df[feat] = hosp_prep.sanitize_diagnoses(df[feat])
if len(diagnosis_feats) > 1:
df = hosp_prep.stack_merger(df)
elif len(diagnosis_feats) == 1:
df.rename(columns={'dx_1': 'cause_code'}, inplace=True)
df['diagnosis_id'] = 1
else:
print("Something went wrong, there are no ICD code features")
start_icds = set(hosp_prep.sanitize_diagnoses(pd.Series(start_icds)))
errors='coerce')
df = df[df.date_adm.notnull()]
df = df[df.date_dis.notnull()]
df['days_diff'] = df.date_dis - df.date_adm
df = df[df.days_diff >= pd.to_timedelta(0, unit="D")]
df = df[(df.days_diff > pd.to_timedelta(0, unit="D")) |
(df.outcome_id == "death")]
diagnosis_feats = df.columns[df.columns.str.startswith('dx_')]
for feat in diagnosis_feats:
df[feat] = hosp_prep.sanitize_diagnoses(df[feat])
if len(diagnosis_feats) > 1:
df = hosp_prep.stack_merger(df)
elif len(diagnosis_feats) == 1:
df.rename(columns={'dx_1': 'cause_code'}, inplace=True)
df['diagnosis_id'] = 1
else:
print("Something went wrong, there are no ICD code features")
df['val'] = 1
print("Are there missing values in any row?\n")
if len(diagnosis_feats) > 1:
# Reshape diagnoses from wide to long
# - review `hosp_prep.py` for additional documentation
df = hosp_prep.stack_merger(df)
df.drop('patient_index', axis=1, inplace=True)
elif len(diagnosis_feats) == 1:
df.rename(columns={'dx_1': 'cause_code'}, inplace=True)
df['diagnosis_id'] = 1
else:
print("Something went wrong, there are no ICD code features")
# 2014 data needs to have the diagnoses cleaned b/c they're good ICD codes.
df.loc[df.source == "GEO_COL_14", 'cause_code'] =\
hosp_prep.sanitize_diagnoses(df.loc[df.source == "GEO_COL_14", 'cause_code'])
#####################################################
# GROUPBY AND AGGREGATE
#####################################################
# Check for missing values
print("Are there missing values in any row?")
null_condition = df.isnull().values.any()
if null_condition:
warnings.warn(">> Yes. ROWS WITH ANY NULL VALUES WILL BE LOST ENTIRELY")
else:
print(">> No.")
group_vars = [
'cause_code', 'diagnosis_id', 'sex_id', 'age_start', 'age_end',
