datatable.countna()

datatable.

countna

(

cols=None

)

sourcetests

Added in version 1.1.0

Count missing values for each column from cols. This function is group-aware.

Parameters

cols

FExpr

Input columns.

return

FExpr

f-expression that counts the number of missing values for each column from cols. If cols is not provided, it will return 0 for each of the frame’s group. The returned f-expression has as many rows as there are groups, it also has the same names and number of columns as in cols. All the resulting column’s stypes are int64.

except

TypeError

The exception is raised when one of the input columns has an obj64 type.

Examples

from datatable import dt, f DT = dt.Frame({'A': [None, 1, 2, None, 2], 'B': [None, 2, 3, 4, 5], 'C': [1, 2, 1, 1, 2]}) DT

	A	B	C
	int32	int32	int32
0	NA	NA	1
1	1	2	2
2	2	3	1
3	NA	4	1
4	2	5	2

5 rows × 3 columns

Count missing values in all the columns:

DT[:, dt.countna(f[:])]

	A	B	C
	int64	int64	int64
0	2	1	0

1 row × 3 columns

Count missing values in the column B only:

DT[:, dt.countna(f.B)]

	B
	int64
0	1

1 row × 1 column

When no cols is passed, this function will always return zero:

DT[:, dt.countna()]

	C0
	int64
0	0

1 row × 1 column

See Also

• count() – function to count the number of non-missing values.