Inequality fails with None and datetime

[amontanez24]

Environment Details

Please indicate the following details about the environment in which you found the bug:

• SDV version: 1.2
• Python version: Any
• Operating System: Any

Error Description

Inequality doesn't work with datetime that has None values

Note that this works if (a) I don't add the constraint or (b) I keep the constraint and I convert the None to pd.NaT

InvalidDataError: The provided data does not match the metadata:
'NoneType' object has no attribute 'to_datetime64'

Steps to reproduce

import pandas as pd
from sdv.metadata import SingleTableMetadata
from sdv.single_table import GaussianCopulaSynthesizer

data = pd.DataFrame(data={
    'A': [None, None, '2020-01-02', '2020-03-04']*2,
    'B': [None, '2021-03-04', '2021-12-31', None]*2
})

metadata = SingleTableMetadata.load_from_dict({
    'columns': {
        'A': { 'sdtype': 'datetime', 'datetime_format': '%Y-%m-%d' },
        'B': { 'sdtype': 'datetime', 'datetime_format': '%Y-%m-%d' }
    }
})

metadata.validate()
synth = GaussianCopulaSynthesizer(metadata)
synth.validate(data)

synth.add_constraints([{
    'constraint_class': 'Inequality',
    'constraint_parameters': {
        'high_column_name': 'A', 
        'low_column_name': 'B'
    }
}])

synth.fit(data)

InvalidDataError: The provided data does not match the metadata:
'NoneType' object has no attribute 'to_datetime64'

[amontanez24]

This can probably be resolved by updating this function in the same way as SDV-Enterprise

SDV/sdv/constraints/utils.py

Lines 10 to 31 in 766d7fc

	def cast_to_datetime64(value):
	"""Cast a given value to a ``numpy.datetime64`` format.
	Args:
	value (pandas.Series, np.ndarray, list, or str):
	Input data to convert to ``numpy.datetime64``.
	Return:
	``numpy.datetime64`` value or values.
	"""
	if isinstance(value, str):
	value = pd.to_datetime(value).to_datetime64()
	elif isinstance(value, pd.Series):
	value.apply(lambda x: pd.to_datetime(x).to_datetime64())
	value = value.astype('datetime64[ns]')
	elif isinstance(value, (np.ndarray, list)):
	value = np.array([
	pd.to_datetime(item).to_datetime64()
	for item in value
	])
	return value