def cardinality(bdb, table, cols=None):
"""Compute the number of unique values in the columns of a table.
Parameters
----------
bdb : __population_to_bdb__
table : __population_name__
Name of table.
cols : list<str>, optional
Columns to compute the unique values. Defaults to all.
Returns
-------
counts : pandas.DataFrame whose .columns are ['name', 'distinct_count'].
"""
# If no columns specified, use all.
if not cols:
sql = 'PRAGMA table_info(%s)' % (quote(table), )
res = bdb.sql_execute(sql)
cols = [r[1] for r in res]
names = []
counts = []
for col in cols:
sql = '''
SELECT COUNT (DISTINCT %s) FROM %s
''' % (quote(col), quote(table))
res = bdb.sql_execute(sql)
names.append(col)
counts.append(cursor_value(res))
return pd.DataFrame({'name': names, 'distinct_count': counts})
def cardinality(bdb, table, cols=None):
"""Compute the number of unique values in the columns of a table.
Parameters
----------
bdb : bayeslite.BayesDB
Active BayesDB instance.
table : str
Name of table.
cols : list<str>, optional
Columns to compute the unique values. Defaults to all.
Returns
-------
counts : list<tuple<str,int>>
A list of tuples of the form [(col_1, cardinality_1), ...]
"""
# If no columns specified, use all.
if not cols:
sql = "PRAGMA table_info(%s)" % (quote(table),)
res = bdb.sql_execute(sql)
cols = [r[1] for r in res]
counts = []
for col in cols:
sql = """
SELECT COUNT (DISTINCT %s) FROM %s
""" % (
quote(col),
quote(table),
)
res = bdb.sql_execute(sql)
counts.append((col, cursor_value(res)))
return counts
def nullify(bdb, table, value):
"""Replace specified values in a SQL table with ``NULL``.
Parameters
----------
bdb : __population_to_bdb__
table : str
The name of the table on which to act
value : stringable
The value to replace with ``NULL``
Examples
--------
>>> import bayeslite
>>> from bdbcontrib import plotutils
>>> with bayeslite.bayesdb_open('mydb.bdb') as bdb:
>>> bdbcontrib.nullify(bdb, 'mytable', 'NaN')
"""
# get a list of columns of the table
c = bdb.sql_execute('pragma table_info({})'.format(quote(table)))
columns = [r[1] for r in c]
for col in columns:
if value in ["''", '""']:
bql = '''
UPDATE {} SET {} = NULL WHERE {} = '';
'''.format(quote(table), quote(col), quote(col))
bdb.sql_execute(bql)
else:
bql = '''
UPDATE {} SET {} = NULL WHERE {} = ?;
'''.format(quote(table), quote(col), quote(col))
bdb.sql_execute(bql, (value, ))
def nullify(bdb, table, value):
"""Replace specified values in a SQL table with ``NULL``.
Parameters
----------
bdb : bayeslite.BayesDB
bayesdb database object
table : str
The name of the table on which to act
value : stringable
The value to replace with ``NULL``
Examples
--------
>>> import bayeslite
>>> from bdbcontrib import plotutils
>>> with bayeslite.bayesdb_open('mydb.bdb') as bdb:
>>> bdbcontrib.nullify(bdb, 'mytable', 'NaN')
"""
# get a list of columns of the table
c = bdb.sql_execute('pragma table_info({})'.format(quote(table)))
columns = [r[1] for r in c]
for col in columns:
if value in ["''", '""']:
bql = '''
UPDATE {} SET {} = NULL WHERE {} = '';
'''.format(quote(table), quote(col), quote(col))
bdb.sql_execute(bql)
else:
bql = '''
UPDATE {} SET {} = NULL WHERE {} = ?;
'''.format(quote(table), quote(col), quote(col))
bdb.sql_execute(bql, (value,))
def get_data_as_list(bdb, table_name, column_list=None):
if column_list is None:
sql = '''
SELECT * FROM {};
'''.format(quote(table_name))
else:
sql = '''
SELECT {} FROM {}
'''.format(', '.join(map(quote, column_list)), table_name)
cursor = bdb.sql_execute(sql)
T = cursor_to_df(cursor).values.tolist()
return T
def get_data_as_list(bdb, table_name, column_list=None):
if column_list is None:
sql = '''
SELECT * FROM {};
'''.format(quote(table_name))
else:
sql = '''
SELECT {} FROM {}
'''.format(', '.join(map(quote, column_list)), table_name)
cursor = bdb.sql_execute(sql)
T = cursor_to_df(cursor).values.tolist()
return T
