How to count null values in postgresql?

Question

select distinct "column" from table;

output:

    column
1     0.0
2     [null]
3     1.0

But when I try to count the null values

select count("column") from train where "column" is NULL;

Gives output 0 (zero)

Can you suggest where it's going wrong?

Answer 1

Use count(*) :

select count(*) from train where "column" is NULL;

count() with any other argument counts the non-NULL values, so there are none if "column" is NULL .

Answer 2

Some workaround when you want to count the values on aggregations, including NULL ones, but can't use count(*) (if other columns are different too).

On these cases, you can use this request :

count(distinct("column")) + (CASE bool_or("column" is null) WHEN true THEN 1 ELSE 0 END)

The count(distinct(column)) will count the non null values, and the other part will add 1 if there is a null value

Answer 3

Use SUM

SELECT SUM(CASE WHEN column IS NULL THEN 1 ELSE 0 END) AS column_null_tally
FROM table;

Answer 4

您得到零是因为您正在计算空（空）值，您需要计算非空字段中的值，例如 id 字段。

select count("id_column") from train where "data_column" is NULL;

Answer 5

Use FILTER<\/code><\/a>

SELECT
  COUNT(*) FILTER (WHERE "column" IS NULL) AS is_null,
  COUNT(*) FILTER (WHERE "column" < 1.0) AS lt_one,
  COUNT(*) FILTER (WHERE "column" > 1.0) AS gt_one,
  COUNT(*) FILTER (WHERE "column" = 1.0) AS just_perfect
FROM "table";

Answer 6

Since

• select count(coalesce(t."column", 0)) from table t is the full number with NULL:s and
• select count(t."column") from table t is the number without NULL:s,

this works as well:

Select count(coalesce(t."column", 0)) - count(t."column") FROM table t;

(This answer might also help those who came here to count both NULL and not NULL, at least I was stranded here since I was searching for that) .