Empty strings and nillable

I have the following:
<xs:element name="AString" nillable="true" type="xs:string" minOccurs="0" maxOccurs="1"/>
What I want is that when I call getAString() that this will return nulll if I
have not set the value.
At the moment it returns an empty string.
Thanks in advance,
Andrew

Hi Andrew,
can you provide a sample?
I tried what you said and I am getting null. Below is what I tried.
Lets say I have an xsd
<?xml version="1.0"?>
<xs:schema
     xmlns:xs="http://www.w3.org/2001/XMLSchema"
     xmlns:bea="http://www.bea.com/xmlbeans/Sample.xsd"
     targetNamespace="http://www.bea.com/xmlbeans/Sample.xsd"
     elementFormDefault="qualified"
     attributeFormDefault="unqualified">
<xs:element name="Person">
<xs:complexType>
<xs:sequence>
<xs:element name="FirstName" type="xs:string" />
<xs:element name="LastName" nillable="true" type="xs:string" minOccurs="0"
maxOccurs="1" />
</xs:sequence>
</xs:complexType>
</xs:element>
</xs:schema>
and below is the code:
PersonDocument personDoc = PersonDocument.Factory.newInstance();
Person person = personDoc.addNewPerson();
person.setFirstName("Mike");
System.out.println("person.getLastName()"+ person.getLastName());
It prints null
Output:
person.getLastName()null
This is on WLW8.1Sp2.
Thanks a lot,
Vimala Ranganathan
"Andrew" <[email protected]> wrote:
>
I have the following:
<xs:element name="AString" nillable="true" type="xs:string" minOccurs="0"
maxOccurs="1"/>
What I want is that when I call getAString() that this will return nulll
if I
have not set the value.
At the moment it returns an empty string.
Thanks in advance,
Andrew

Similar Messages

How to check empty string and null? Assign same value to multiple variables

Hi,
1.
How do I check for empty string and null?
in_value IN VARCHAR2
2. Also how do I assign same value to multiple variables?
var_one NUMBER := 0;
var_two NUMBER := 0;
var_one := var_two := 0; --- Gives an error
Thanks

MichaelS wrote:
Not always: Beware of CHAR's:
Bug 727361: ZERO-LENGTH STRING DOES NOT RETURN NULL WHEN USED WITH CHAR DATA TYPE IN PL/SQL:
SQL> declare
2    l_str1   char (10) := '';
3    l_str2   char (10) := null;
4 begin
5
6    if l_str1 is null
7    then
8      dbms_output.put_line ('oh STR1 is null');
9    elsif l_str1 is not null
10    then
11      dbms_output.put_line ('oh STR1 is NOT null');
12    end if;
13
14    if l_str2 is null
15    then
16      dbms_output.put_line ('oh STR2 is null');
17    elsif l_str2 is not null
18    then
19      dbms_output.put_line ('oh STR2 is NOT null');
20    end if;
21 end;
22 /
oh STR1 is NOT null
oh STR2 is null
PL/SQL procedure successfully completed.
SQL> alter session set events '10932 trace name context forever, level 16384';
Session altered.
SQL> declare
2    l_str1   char (10) := '';
3    l_str2   char (10) := null;
4 begin
5
6    if l_str1 is null
7    then
8      dbms_output.put_line ('oh STR1 is null');
9    elsif l_str1 is not null
10    then
11      dbms_output.put_line ('oh STR1 is NOT null');
12    end if;
13
14    if l_str2 is null
15    then
16      dbms_output.put_line ('oh STR2 is null');
17    elsif l_str2 is not null
18    then
19      dbms_output.put_line ('oh STR2 is NOT null');
20    end if;
21 end;
22 /
oh STR1 is null
oh STR2 is null
PL/SQL procedure successfully completed.
SQL> SY.

Varchar2, empty strings and NULL

Hi all,
When inserting an empty string into a column of type varchar2 - is a NULL value stored in the column by the database? I've seen conflicting reports, and I know that the SQL 1992 spec specifies that empty strings not be treated as a NULL value, but that Oracle has traditionally treated zero length strings stored in a varchar2 column as NULL.
So, is there a way to store an empty string in a varchar2 column as an empty string and not a NULL value?
TIA,
Seth

It can be even more complicated or annoying than NULL not equal to ASCII NULL.
example:
create table test_null
(fld1 varchar2(10),
fld2 varchar2(10),
fld3 varchar2(20));
insert into test_null values (chr(0),null, 'chr(0) and null');
insert into test_null values (null, null, 'null and null');
insert into test_null values ('', chr(0), ''''' and chr(0)');
insert into test_null values ('', null, ''''' and null');
select * from test_null;
FLD1       FLD2       FLD3
                      chr(0) and null
                      null and null
                      '' and chr(0)
                      '' and null
1 DECLARE
2 BEGIN
3   for c1 in (select fld1, fld2, fld3 from test_null) loop
4      if c1.fld1 = c1.fld2 then
5         dbms_output.put_line(c1.fld3||' Are equal'||
6                ' Length fld1 = '||to_char(length(c1.fld1))||
7                ' Length fld2 = '||to_char(length(c1.fld2)));
8      else
9         dbms_output.put_line(c1.fld3||' Are NOT equal'||
10                ' Length fld1 = '||to_char(length(c1.fld1))||
11                ' Length fld2 = '||to_char(length(c1.fld2)));
12      end if;
13      dbms_output.put_line(' ');
14   end loop;
15* END;
SQL> /
chr(0) and null Are NOT equal Length fld1 = 1 Length fld2 =
null and null Are NOT equal Length fld1 = Length fld2 =
'' and chr(0) Are NOT equal Length fld1 = Length fld2 = 1
'' and null Are NOT equal Length fld1 = Length fld2 =
PL/SQL procedure successfully completed.

Distinguishing between empty string and null values

hi all,
I am using an ODBC connection to connect my java app to database using JDBCODBC driver. I have 2 columns 'aColumn' and 'bColumn' in my table, both allow null values. I have one row in it, which has null value in aColumn and empty string in bColumn. I retrieve this row's data and assign it to 2 columns : 'aColumnVar' and 'bColumnVar' respectively. I find out that both 'aColumnVar' and 'bColumnVar' variables has null values (although bColumnVar should has an empty string as its value). Now my ODBC connection Data Source has the option "Use ANSI nulls, paddings, and warnings" ON. I turn it off and try again. This time both 'aColumnVar' and 'bColumnVar' variables has empty string as values (although aColumnVar should has null as its value).
How can I make sure that i can get the data exactly as it is in the database in my variables?
Thanks

there is a wasNull() method on ResultSet. After you
have obtained the value of a column e.g. by calling a
method like getString you can call wasNull and if it
returns true then the value on the database is null.
Check the java docs, it might explain it better
http://java.sun.com/j2se/1.4.1/docs/api/java/sql/Result
et.html#wasNull()I am using MS SQL Server 7.0 under Windows NT 4.0 with JDK 1.2. My ODBC connection Data Source has to have the option "Use ANSI nulls, paddings, and warnings" ON.
I try the wasNull() method but it is doing the same thing i.e. telling me that a column is null when in database it is null (right); and a column is null when in database it is an empty string (wrong). I suspect it is something to do with the JDBC-ODBC driver I am using.

Empty string and NULL value

Hi,
Does oracle treat empty string as NULL.
If it is so,Can we force db to treat both as different things.
Thanx

11g.DBA wrote:
Hi,
Does oracle treat empty string as NULL.Yes.
If it is so,Can we force db to treat both as different things.No.
Please read in SQL Reference:
http://docs.oracle.com/cd/E11882_01/server.112/e26088/sql_elements005.htm#SQLRF3003

Is empty string and Null same?

create table x
(empid number);
Table created.
insert into x values ('');
1 row created.
insert into x values (null);
1 row created.
SQL> set null <<>>
SQL> select * from x;
     EMPID
<<>>
<<>>So i can safely change an Insert statement like
insert into x values ('');to
insert into x values (null);Right?

michaels2 wrote:
char pads with spaces to the length of the data itemdon't think so - at least not on my 11.2.0.1.0Good. Oracle finally fixed it. Before 11.2 it was pretty much as ajallen noted:
Oracle Database 11g Enterprise Edition Release 11.1.0.6.0 - Production
PL/SQL Release 11.1.0.6.0 - Production
CORE    11.1.0.6.0      Production
TNS for 32-bit Windows: Version 11.1.0.6.0 - Production
NLSRTL Version 11.1.0.6.0 - Production
SET SERVEROUTPUT ON
DECLARE
    flag CHAR(2);
    PROCEDURE check_null (p_flag IN CHAR)
    IS
    BEGIN
      IF p_flag = ' '
        THEN
          dbms_output.put_line ('flag is equal to '' ''');
        ELSIF p_flag IS NULL
          THEN
            dbms_output.put_line ('flag is null');
        ELSE
          dbms_output.put_line ('other');
      END IF;
    END;
BEGIN
    flag := '';
    check_null (flag);
    flag := NULL;
    check_null (flag);
END;
flag is equal to ' '
flag is null
PL/SQL procedure successfully completed.
alter session set events '10932 trace name context forever, level 16384';
DECLARE
    flag CHAR(2);
    PROCEDURE check_null (p_flag IN CHAR)
    IS
    BEGIN
      IF p_flag = ' '
        THEN
          dbms_output.put_line ('flag is equal to '' ''');
        ELSIF p_flag IS NULL
          THEN
            dbms_output.put_line ('flag is null');
        ELSE
          dbms_output.put_line ('other');
      END IF;
    END;
BEGIN
    flag := '';
    check_null (flag);
    flag := NULL;
    check_null (flag);
END;
flag is null
flag is null
PL/SQL procedure successfully completed.
SQL>
Bug 727361: ZERO-LENGTH STRING DOES NOT RETURN NULL WHEN USED WITH CHAR DATA TYPE IN PL/SQL
SY.

Null and empty string not being the same in object?

Hello,
I know that null and empty string are interpreted the same in oracle.
However I discovered the strange behaviour concerning user defined objects:
create or replace
TYPE object AS OBJECT (
value VARCHAR2(2000)
declare
xml xmltype;
obj object;
begin
obj := object('abcd');
xml := xmltype(obj);
dbms_output.put_line(xml.getStringVal());
obj.value := '';
xml := xmltype(obj);
dbms_output.put_line(xml.getStringVal());
obj.value := null;
xml := xmltype(obj);
dbms_output.put_line(xml.getStringVal());
end;
When creating xml from object, all not-null fields are transformed into xml tag.
I supposed that obj.value being either '' or null will lead to the same result.
However this is output from Oracle 9i:
<OBJECT_ID><VALUE>abcd</VALUE></OBJECT_ID>
<OBJECT_ID><VALUE></VALUE></OBJECT_ID>
<OBJECT_ID/>
Oracle 10g behaves as expected:
<OBJECT><VALUE>abcd</VALUE></OBJECT>
<OBJECT/>
<OBJECT/>
However Oracle 9i behaviour leads me to the conclusion that oracle
must somehow distinguish between empty string and null in user defined objects...
Can someone clarify this behaviour?
Thus is it possible to test if object's field is empty or null?

However Oracle 9i behaviour leads me to the conclusion that oracle
must somehow distinguish between empty string and null in user defined objects...
Can someone clarify this behaviour?
Thus is it possible to test if object's field is empty or null?A lot of "fixes" were done, relating to XML in 10g and the XML functionality of 9i was known to be buggy.
I think you can safely assume that null and empty strings are treated the same by Oracle regardless. If you're using anything less than 10g, it's not supported any more anyway, so upgrade. Don't rely on any assumptions that may appear due to bugs.

StringTokenizer vs. split and empty strings -- some clarification please?

Hi everybody,
I posted a question that was sort of similar to this once, asking if it was best to convert any StringTokenizers to calls to split when parsing strings, but this one is a little different. I rarely use split, because if there are consecutive delimiters, it gives empty strings in the array it returns, which I don't want. On the other hand, I know StringTokenizer is slower, but it doesn't give empty strings with consecutive delimiters. I would use split much more often if there was a way to use it and not have to check every array element to make sure it isn't the empty string. I think I may have misunderstood the javadoc to some extent--could anyone explain to me why split causes empty strings and StringTokenizer doesn't?
Thanks,
Jezzica85

Because they are different?
Tokenizers are designed to return tokens, whereas split is simply splitting the String up into bits. They have different purposes
and uses to be honest. I believe the results of previous discussions of this have indicated that Tokenizers are slightly (very
slightly and not really meaningfully) faster and tokenizers do have the option of return delimiters as well which can be useful
and is a functionality not present in just a straight split.
However. split and regex in general are newer additions to the Java platform and they do have some advantages. The most
obvious being that you cannot use a tokenizer to split up values where the delimiter is multiple characters and you can with
split.
So in general the advice given to you was good, because split gives you more flexibility down the road. If you don't want
the empty strings then yes just read them and throw them away.
Edited by: cotton.m on Mar 6, 2008 7:34 AM
goddamned stupid forum formatting

How to distinguish NULL and Empty Strings

Hi,
Just to set the context right; I'm an experienced C programmer trying labview for the first time. As such I ran in to a problem being that Labview has no concept of NULL-pointers and more specifically appears to have no concept of the difference between a NULL-string and an empty-string
I'm trying to make a structure (bundle) of strings (in it's most basic form key-value pairs) which i'd like to (for instance) URI encode in order to send it to a web server. For those who are not familiar with URI encoding; there is a distinguished difference between setting a key to an empty string and setting a key with no value. In C I would use a pointer to an empty string vs a NULL string pointer to symbolize this.
In essence I need an elegant way to distinguish between a defined but empty string and an undefined string (hmmm this is actually describing the same problem but now in terms of perl).
Anybody have any pointers (pun not intended) for me ?

This is a bit depending on the interface you have with your encoder. The whole issue is that LV has no pointers at all (and you will like it, as you will never have any Null-Pointer exeptions and the like).
Assuming that you use a dll (so the Call Library node).
Use CString as input -> NULL-Terminated String.
Use I32 as input and pass 0 -> NULL string.
Felix
www.aescusoft.de
My latest community nugget on producer/consumer design
My current blog: A journey through uml

Setting input fields to an empty string BIG PROBLEM

Hi Rob, thanks for your rerply, It makes sense alright, but I am using 2 seperate input fields. I have discovered what is causing the problem .
After the user types into the inputfield and moves on to a different frame, I have set the input field to an empty string and this is what is causing the problem.
I have tried setting the input field to an empty string in a movie script and on individual sprites and the first right answer is not excepted.I have to type in the right answer twice to get get a right response. I am actually using three seperate input fields in total, each with a different name and behavior. Even If I use one input field I still have the same problem. Its driving me crazey. Any Ideas.
Anne

I believe that the FIM Service always does trims on string, so you may be out of luck.
What is your scenario / what are you trying to accomplish by setting a space in an attribute? A space is not an empty string in my definition.
Regards, Soren Granfeldt
blog is at http://blog.goverco.com | facebook https://www.facebook.com/TheIdentityManagementExplorer | twitter at https://twitter.com/#!/MrGranfeldt

SQLLDR - how to set a default value for an empty string

I'm using SQLLDR to load data from a flat file on Oracle 10g 10.2.0.4 EE, Linux RHEL 4 update 6. Seems like my question is simple, but darned if I can't find the solution.
My source file (comma delimited extracted form a different RDBMS) has some (not all) rows where the output for a string is "", and not NULL. How do I check for empty string and add a default value in the control file? The target column is non-nullable
Here's my current control file - however, the column attributeName where sometimes the source record is "" does not get evaluated as NULL.
load data
infile 'mydata.del'
badfile 'mydata.bad'
discardfile 'mydata.dsc'
into table mydata
fields terminated by "," optionally enclosed by '"'
   TRAILING NULLCOLS
(l_metadata,
company,
className,
attributeName "decode(:vc_attributeName,null, 'none')",
attributeDefault)
{code}
When I tried decode with "", I received the following error:
{code}SQL*Loader-350: Syntax error at line 11.
Expecting "," or ")", found ", 'none')".
attributeName "decode(:vc_attributeName,"", 'none')", {code}
This does not work either:
{code}
load data
infile 'mydata.del'
badfile 'mydata.bad'
discardfile 'mydata.dsc'
into table mydata
fields terminated by "," optionally enclosed by '"'
   TRAILING NULLCOLS
(l_metadata,
company,
className,
attributeName NULLIF attributeName = BLANKS ,"decode(:vc_attributeName,null, 'none')",
attributeDefault) I get the following in the log:
Table COMPANYMETADATA, loaded from every logical record.
Insert option in effect for this table: INSERT
TRAILING NULLCOLS option in effect
   Column Name                  Position   Len Term Encl Datatype
L_METADATA                          FIRST     *   , O(") CHARACTER
COMPANY                            NEXT     *   , O(") CHARACTER
CLASSNAME                         NEXT     *   , O(") CHARACTER
ATTRIBUTENAME                     NEXT     *   , O(") CHARACTER
    NULL if ATTRIBUTENAME = BLANKS
    SQL string for column : "decode(:attributeName,null, 'none')"
ATTRIBUTEDEFAULT                  NEXT     *   , O(") CHARACTER
Record 1: Rejected - Error on table COMPANYMETADATA, column ATTRIBUTENAME.
ORA-01400: cannot insert NULL into ("SRV5"."COMPANYMETADATA"."ATTRIBUTENAME")
{code}
Any help is appreciated -
Edited by: kpw on Feb 9, 2009 11:10 AM

Hello,
I just loaded following data successfully in my_data table using the same control file I posted. Yes it will evaluate "" as null and replace with none, remember in control file you have this option, "optionally enclosed by '"'.
#my_data.dat
"myname1", "attribname1"
"myname2", "attribname2"
"myname3", "attribname3"
"myname4", ""
#my_data.ctl
load data
into table my_data
fields terminated by "," optionally enclosed by '"'
TRAILING NULLCOLS
my_name char(30),
vc_attributeName "decode(:vc_attributeName,null, 'none', :vc_attributeName)"
{code}
Here is the output log file
{code}
Control File:   my_data.ctl
Data File:      mydata.dat
Bad File:     mydata.bad
Discard File: none specified
(Allow all discards)
Number to load: ALL
Number to skip: 0
Errors allowed: 50
Bind array:     64 rows, maximum of 256000 bytes
Continuation:    none specified
Path used:      Conventional
Table MY_DATA, loaded from every logical record.
Insert option in effect for this table: INSERT
TRAILING NULLCOLS option in effect
   Column Name                  Position   Len Term Encl Datatype
MY_NAME                             FIRST    30   , O(") CHARACTER
VC_ATTRIBUTENAME                     NEXT     *   , O(") CHARACTER
    SQL string for column : "decode(:vc_attributeName,null, 'none', :vc_attributeName)"
Table MY_DATA:
4 Rows successfully loaded.
0 Rows not loaded due to data errors.
0 Rows not loaded because all WHEN clauses were failed.
0 Rows not loaded because all fields were null.
Space allocated for bind array:                  18560 bytes(64 rows)
Read   buffer bytes: 1048576
Total logical records skipped:          0
Total logical records read:             4
Total logical records rejected:         0
Total logical records discarded:        0
Run began on Mon Feb 09 18:17:35 2009
Run ended on Mon Feb 09 18:17:36 2009
Elapsed time was:     00:00:00.98
CPU time was:         00:00:00.10
{code}
Hope this helps clearing your dobuts
Regards

Problem: textbox leave empty string("") instead of null

Hello, I have a problem with textboxes. By default, they always are initialized to null, but when I start entering and removing data from textboxes, many times they leave an empty string "" instead of null.
And, of course, it's not the same a null value as an empty string and thus, I haven't got the expected result in queries to database.
Thus, I always have to compare textbox values with "" and null. It's very tedious and not elegant solution. I would like to know if there is another way to get the values as null instead of empty strings.

Yes. Once you entered and remove the text it will evaluated as you told .
For ur case u can Try the condition as
if ( instance !=null && instance.length !=0)
be sure instance != null check b4 other wise u can get NullPointerException

Empty string == NULL ???!?!?!?!??!

Hi SQlers ...
I'm using Oracle 8.1.7 ... and Oracle maps the literal empty
string value, to the NULL value ... it that the expected result?
simple example:
SELECT 1 from dual where ('' is null)
I've tried the same on others rdbms, and of course, the
result is not the same.
thanks in advance
saludos
dario estepario ...

Yup. In Oracle the empty string '' and NULL are identical, though they periodically threaten that this may change in the future.
Justin

Oracle, Null and empty Strings

Currently I'm facing problems with a class, which contains a String, which
is set to "" (empty String).
When the class is persistent, oracle writes null to the table column
(which seems to be common oracle behaviour) and when retrieving the class,
the field is set to null as well, giving me a lot of null-pointer
exceptions.
Anyway ... I can cope with that (just a lot of extra work)
far worse is the problem, wenn searching objects, that have this field set
to "" oder null.
Oracle can't find the records because JDO creates Querys "where
string=null" or "where string=''" , where oracle expects "where string is
null" to find the records.
Is there a workaround or solution ?

Yeah, that would work as well, thx, but since I have to cope with
null-Strings now everywhere in my program, it doesn't hurt just to forbid
empty strings on the program side.
In future times I'll test on Oracle first, then porting to DB/2 - this way
I suppose work is far less to garant compability.
Nevertheless ... having to set the bankcode into quotes is a kodo bug in
my opinion.
Kodo knows the type of classfields (in this case string) and shouldn't
send the parameter as a BigDecimal to the database.
Given that, and having only bankcodes of null (only neccesary when using
Oracle), the method would look like:
public Collection getAccounts (String bankCode)
throws Exception
return getAccounts (Account.class, "bankcode=="+bankcode);
which is how a transparent persistent layer, um, should be , um , I mean
... just transparent ;-D
Marc Prud'hommeaux wrote:
Stefan-
Couldn't you just do something like:
public Collection getAccounts (String bankCode)
throws Exception
String filter;
if (bankCode == null || bankCode.length () == 0)
filter = "(bankCode == null || bankCode == "")";
else
filter = "bankCode == "" + bankCode + """;
return getAccounts (Account.class, filter);
If I understand the problem correctly, this should work for all the
databases.
In article <[email protected]>, Stefan wrote:
What operations are you performing to cause this SQL to be issued? You
say you are having trouble removing objects, but this is clearly not a
DELETE statement. Is this the SQL that is issued when looking up
objects by identity?I'm not removing objects, I was removing just quotes from parameters ;-)
A string column... is it also represented as a string field in your class?Yeah ... just to give you an impression of the code:
First we have a class, representing a bank account:
public class Account {
private AccountMgr myAccountMgr;
private String bankCode;
private String id;
Note, that in nearly all cases bankCode will be a number or null.
I have a second class "AccountMgr", which does all of the persistant stuff
(seaching, making persistent etc.)
This class has two methods, one versatile (protected) to retrieve accounts
by a given filterString and one who just returns accounts by bankCode,
building the expected filterstring. Here is my current working version:
public class AccountMgr {
public Collection getAccounts(String bankCode) throws Exception {
if (bankCode!=null) {
if (bankCode.equals("")) {
throw new Exception("check code, bankCode='' not allowed to get
same behavior from DB2 and Oracle");
// if set, quote the bankCode
bankCode="""+bankCode+""";
return getAccounts(Account.class,"bankCode=="+bankCode);
protected Collection getAccounts(Class accountClass, String filterAdd)
throws Exception {
PersistenceManager pm = MyHelper.getPersistenceManager();
String filter="";
if (filterAdd!=null && !filterAdd.trim().equals("")) {
filter+=filterAdd + " && ";
filter += "myAccountMgr==_accMgr";
Query query = pm.newQuery(accountClass, filter);
query.declareParameters("AccountMgr _accMgr");
return (Collection) query.execute(this);
As you can see, in the first method I have to set the bankCode into
quotes, when it's not null.
This is because otherwise a filter like "bankCode=1234" will be translated
in a way, where 1234 is send as a BigDecimal to the database:
[...] executing statement <4239745>: (SELECT [...] FROM JDO_ACCOUNT t0
WHERE t0.BANKCODE = ? : [reused=1;params={(BigDecimal) 1234}]
Marc Prud'hommeaux [email protected]
SolarMetric Inc. http://www.solarmetric.com

NULL and Empty String

Hi There,
As far as I know, Null is not the same as an empty string; however, when I try this out, I get some unexpected results (well, at least unexpected for my liking):
SQL> CREATE TABLE TS (MID NUMBER,
2 MDESC VARCHAR2(20) DEFAULT '' NOT NULL);
Table created.
SQL> INSERT INTO TS VALUES(1,'');
INSERT INTO TS VALUES(1,'')
ERROR at line 1:
ORA-01400: cannot insert NULL into ("TT"."TS"."MDESC")So, according to the above scenario, I can't insert an empty string!! However, an empty string is a valid string that doesn't have tuples/data!!
How come Oracle translates the null string '' as NULL?
Thanks

William Robertson wrote:
There is a special case to do with CHAR values, whereby '' counts as a string and so gets blank-padded, whereas NULL does not.Are you referring to:
SQL> DECLARE
2      flag CHAR(2);
3      PROCEDURE check_null (p_flag IN CHAR)
4      IS
5      BEGIN
6        IF p_flag = ' '
7        THEN
8          dbms_output.put_line ('flag is equal to '' ''');
9        ELSIF p_flag IS NULL
10        THEN
11          dbms_output.put_line ('flag is null');
12        ELSE
13          dbms_output.put_line ('other');
14        END IF;
15      END;
16    BEGIN
17      flag := '';
18      check_null (flag);
19      flag := NULL;
20      check_null (flag);
21    end;
22 /
flag is equal to ' '
flag is null
PL/SQL procedure successfully completed.
SQL> alter session set events '10932 trace name context forever, level 16384';
Session altered.
SQL> DECLARE
2      flag CHAR(2);
3      PROCEDURE check_null (p_flag IN CHAR)
4      IS
5      BEGIN
6        IF p_flag = ' '
7        THEN
8          dbms_output.put_line ('flag is equal to '' ''');
9        ELSIF p_flag IS NULL
10        THEN
11          dbms_output.put_line ('flag is null');
12        ELSE
13          dbms_output.put_line ('other');
14        END IF;
15      END;
16    BEGIN
17      flag := '';
18      check_null (flag);
19      flag := NULL;
20      check_null (flag);
21    end;
22 /
flag is null
flag is null
PL/SQL procedure successfully completed.
SQL> SY.
P.S. Don't ask me why normal (or at least consistent) behavior is not the default.

Empty strings and nillable

Similar Messages

Maybe you are looking for