The BIT and VARBIT data types stores bit arrays.
With BIT, the length is fixed; with VARBIT, the length can be variable.
Aliases
The name BIT VARYING is an alias for VARBIT.
Syntax
Bit array constants are expressed as literals. For example, B'100101' denotes an array of 6 bits.
For more information about bit array constants, see the constants documentation on bit array literals.
For usage, see the Example below.
Size
The number of bits in a BIT value is determined as follows:
| Type declaration | Logical size |
|---|---|
| BIT | 1 bit |
| BIT(N) | N bits |
| VARBIT | variable with no maximum |
| VARBIT(N) | variable with a maximum of N bits |
The effective size of a BIT value is larger than its logical number
of bits by a bounded constant factor. Internally, CockroachDB stores
bit arrays in increments of 64 bits plus an extra integer value to
encode the length.
The total size of a BIT value can be arbitrarily large, but it is
recommended to keep values under 1 MB to ensure performance. Above
that threshold, write
amplification and
other considerations may cause significant performance degradation.
Example
> CREATE TABLE b (x BIT, y BIT(3), z VARBIT, w VARBIT(3));
> SHOW COLUMNS FROM b;
column_name | data_type | is_nullable | column_default | generation_expression | indices | is_hidden
+-------------+-----------+-------------+----------------+-----------------------+-----------+-----------+
x | BIT | true | NULL | | {primary} | false
y | BIT(3) | true | NULL | | {primary} | false
z | VARBIT | true | NULL | | {primary} | false
w | VARBIT(3) | true | NULL | | {primary} | false
rowid | INT | false | unique_rowid() | | {primary} | true
> INSERT INTO b(x, y, z, w) VALUES (B'1', B'101', B'1', B'1');
> SELECT * FROM b;
x | y | z | w
+---+-----+---+---+
1 | 101 | 1 | 1
For type BIT, the value must match exactly the specified size:
> INSERT INTO b(x) VALUES (B'101');
pq: bit string length 3 does not match type BIT
> INSERT INTO b(y) VALUES (B'10');
pq: bit string length 2 does not match type BIT(3)
For type VARBIT, the value must not be larger than the specified maximum size:
> INSERT INTO b(w) VALUES (B'1010');
pq: bit string length 4 too large for type VARBIT(3)
Supported casting and conversion
BIT values can be cast to any of the following data types:
| Type | Details |
|---|---|
INT |
Converts the bit array to the corresponding numeric value, interpreting the bits as if the value was encoded using two's complement. If the bit array is larger than the integer type, excess bits on the left are ignored. For example, B'1010'::INT equals 10. |
STRING |
Prints out the binary digits as a string. This recovers the literal representation. For example, B'1010'::STRING equals '1010'. |
See also
Was this helpful?
