Fix representation of categorical data types in NewRowSynthesis metric

[MaanasArora]

Resolves #397.

Reviewing the history of the file, it seems that the condition of checking for the O (object) data type was created to hard-code single quotes around the value instead of using repr:

SDMetrics/sdmetrics/single_table/new_row_synthesis.py

Lines 120 to 123 in 585290f

	if real_data[field].dtype == 'O':
	field_filter = f'`{field}` == {repr(row[field])}'
	else:
	field_filter = f'`{field}` == {row[field]}'

But now that we have it, I propose we can remove the else condition as repr would handle our edge cases (such as formatting strings with quotes and integers with none).

Let me know if this works. Thank you very much!