mirror of
https://github.com/apache/sqoop.git
synced 2025-05-04 17:12:35 +08:00
SQOOP-962: Document null string behavior
(Venkat Ranganathan via Jarek Jarcec Cecho)
This commit is contained in:
parent
a0a03d7f31
commit
7048ca3023
@ -310,3 +310,59 @@ $ sqoop export \
|
|||||||
--export-dir hdfsdir \
|
--export-dir hdfsdir \
|
||||||
--input-fields-terminated-by "\t"
|
--input-fields-terminated-by "\t"
|
||||||
----
|
----
|
||||||
|
|
||||||
|
Null string handling
|
||||||
|
^^^^^^^^^^^^^^^^^^^^
|
||||||
|
|
||||||
|
Netezza direct connector supports the null-string features of Sqoop. The null
|
||||||
|
string values are converted to appropriate external table options during export
|
||||||
|
and import operations.
|
||||||
|
|
||||||
|
.Supported export control arguments:
|
||||||
|
[grid="all"]
|
||||||
|
`----------------------------------------`---------------------------------------
|
||||||
|
Argument Description
|
||||||
|
---------------------------------------------------------------------------------
|
||||||
|
+\--input-null-string <null-string>+ The string to be interpreted as\
|
||||||
|
null for string columns.
|
||||||
|
+\--input-null-non-string <null-string>+ The string to be interpreted as\
|
||||||
|
null for non string columns.
|
||||||
|
---------------------------------------------------------------------------------
|
||||||
|
|
||||||
|
In the case of Netezza direct mode connector, both the arguments must be
|
||||||
|
left to the default values or explicitly set to the same value. Furthermore
|
||||||
|
the null string value is restricted to 0-4 utf8 characters.
|
||||||
|
|
||||||
|
On export, for non-string columns, if the chosen null value is a valid
|
||||||
|
representation in the column domain, then the column might not be loaded as
|
||||||
|
null. For example, if the null string value is specified as "1", then on
|
||||||
|
export, any occurrence of "1" in the input file will be loaded as value 1
|
||||||
|
instead of NULL for int columns.
|
||||||
|
|
||||||
|
It is suggested that the null value be specified as empty string for
|
||||||
|
performance and consistency.
|
||||||
|
|
||||||
|
.Supported import control arguments:
|
||||||
|
[grid="all"]
|
||||||
|
`----------------------------------------`---------------------------------------
|
||||||
|
Argument Description
|
||||||
|
---------------------------------------------------------------------------------
|
||||||
|
+\--null-string <null-string>+ The string to be interpreted as\
|
||||||
|
null for string columns.
|
||||||
|
+\--null-non-string <null-string>+ The string to be interpreted as\
|
||||||
|
null for non string columns.
|
||||||
|
---------------------------------------------------------------------------------
|
||||||
|
|
||||||
|
In the case of Netezza direct mode connector, both the arguments must be
|
||||||
|
left to the default values or explicitly set to the same value. Furthermore
|
||||||
|
the null string value is restricted to 0-4 utf8 characters.
|
||||||
|
|
||||||
|
On import, for non-string columns, the chosen null value in current
|
||||||
|
implementations the null value representation is ignored for non character
|
||||||
|
columns. For example, if the null string value is specified as "\N", then on
|
||||||
|
import, any occurrence of NULL for non-char columns in the table will be
|
||||||
|
imported as an empty string instead of '\N', the chosen null string
|
||||||
|
representation.
|
||||||
|
|
||||||
|
It is suggested that the null value be specified as empty string for
|
||||||
|
performance and consistency.
|
||||||
|
Loading…
Reference in New Issue
Block a user