I have a Postgres table with a string column carrying numeric values. I need to convert these strings to numbers for math, but I need both NULL
values as well as empty strings to be interpreted as 0
.
I can convert empty strings into null values:
# select nullif('','');
nullif
--------
(1 row)
And I can convert null values into a 0
:
# select coalesce(NULL,0);
coalesce
----------
0
(1 row)
And I can convert strings into numbers:
# select cast('3' as float);
float8
--------
3
(1 row)
But when I try to combine these techniques, I get errors:
# select cast( nullif( coalesce('',0), '') as float);
ERROR: invalid input syntax for integer: ""
LINE 1: select cast( nullif( coalesce('',0), '') as float);
# select coalesce(nullif('3',''),4) as hi;
ERROR: COALESCE types text and integer cannot be matched
LINE 1: select coalesce(nullif('3',''),4) as hi;
What am I doing wrong?
Its just a literal. Null is the value of reference variable. But empty string is blank.It gives the length=0 . Empty string is a blank value,means the string does not have any thing.
Multiplying a string by 1 ( [string] * 1 ) works just like the unary plus operator above, it converts the string to an integer or floating point number, or returns NaN (Not a Number) if the input value doesn't represent a number. Like the Number() function, it returns zero ( 0 ) for an empty string ( '' * 1 ).
The Java programming language distinguishes between null and empty strings. An empty string is a string instance of zero length, whereas a null string has no value at all. An empty string is represented as "" . It is a character sequence of zero characters.
There are two ways to replace NULL with blank values in SQL Server, function ISNULL(), and COALESCE(). Both functions replace the value you provide when the argument is NULL like ISNULL(column, '') will return empty String if the column value is NULL.
The types of values need to be consistent; coalescing the empty string to a 0 means that you cannot then compare it to null
in the nullif
. So either of these works:
# create table tests (orig varchar);
CREATE TABLE
# insert into tests (orig) values ('1'), (''), (NULL), ('0');
INSERT 0 4
# select orig, cast(coalesce(nullif(orig,''),'0') as float) as result from tests;
orig | result
------+--------
1 | 1
| 0
| 0
0 | 0
(4 rows)
# select orig, coalesce(cast(nullif(orig,'') as float),0) as result from tests;
orig | result
------+--------
1 | 1
| 0
| 0
0 | 0
(4 rows)
You could also use
cast(
case
when coalesce(orig, '') = '' then '0'
else orig
end
as float
)
You could also unwrap that a bit since you're being fairly verbose anyway:
cast(
case
when orig is null then '0'
when orig = '' then '0'
else orig
end
as float
)
or you could put the cast inside the CASE:
case
when coalesce(orig, '') = '' then 0.0
else cast(orig as float)
end
A CASE makes it a bit easier to account for any other special conditions, this also seems like a clearer expression of the logic IMO. OTOH, personal taste and all that.
Actually, you can cast NULL to int, you just can't cast an empty string to int. Assuming you want NULL in the new column if data1 contains an empty string or NULL, you can do something like this:
UPDATE table SET data2 = cast(nullif(data1, '') AS int);
or
UPDATE table SET data2 = nullif(data1, '')::int;
Reference
If you love us? You can donate to us via Paypal or buy me a coffee so we can maintain and grow! Thank you!
Donate Us With