How to handle blank values in FOR XML PATH/AUTO - SQL Server? - sql-server

Does anybody have any insight or encountered with this SQL Server behavior,
when blank value is converted into data type "char" or column which have "char" as data type - processed using For XML PATH, it returned XML result with "" special characters encoding for space,
When same blank value converted with varchar - it returns empty tag. My desired result is empty tag.
SELECT field=CONVERT(CHAR(10),'')
FOR XML PATH(''), ELEMENTS
--RESULT: <field> </field>
SELECT field=CONVERT(VARCHAR(10),'')
FOR XML PATH(''), ELEMENTS
--RESULT: <field></field>
The explanation in my view is when I'm using char it is inserting 10 spaces. And ASCII Hex20 or Decimal 32 - is a code for space.
It can be handled by using varchar instead of char as data type. But in my case I am reading field value from table defined in database:
--Example:
CREATE TABLE #temp(field CHAR(2))
INSERT INTO #temp
SELECT NULL
SELECT field=ISNULL(field,'')
FROM #temp
FOR XML PATH('')
DROP TABLE #temp
--RESULT: <field> </field>
--Desired Result:<field></field>
What is a most elegant way/or possibilities in SQL?
PS: I have a result set of 50 columns. I would like to handle this at database level without changing types. Real problem is my web service when encounter this character in XML throws a potentially dangerous value error.

Solution I used for desired result is, casting a whole result at once -
CREATE TABLE #temp(field CHAR(2))
INSERT INTO #temp
SELECT NULL
--Adding a cast to XML Result to convert it into varchar & then use replace.
SELECT REPLACE(CONVERT(VARCHAR(max),(SELECT field=ISNULL(field,'')
FROM #temp
FOR XML PATH(''))),' ','')
DROP TABLE #temp
would like to hear if there is any other suggestion?

You could do the following...up to you if you feel you need to TRIM or not;
CREATE TABLE #temp(field CHAR(2))
INSERT INTO #temp
SELECT NULL
SELECT field=COALESCE(NULLIF(RTRIM(LTRIM(field)), ''), '')
FROM #temp
FOR XML PATH('')
DROP TABLE #temp

Related

How does this query work to split csv into rows in sql server?

I am using sql-server 2012
here is the query
drop table #t
create table #t(id int,name varchar(10))
insert into #t values(1,'a,b,c'),(2,'d,e')
select A.id,split.a.value('.','varchar(10)') as string
from
(select id,cast('<M>'+replace([name],',','</M><M>')+'</M>' as xml) as string from #t) as A cross apply string.nodes('/M') as split(a)
I am trying to get the logic of query used here,but unable to understand how it is working?
can some body guide me in making understand the
logic in the cast() clause and split.a.value()
The Cast clause converts the string into an XML fragment to be consumed later by the 'nodes' method. If you run just the subquery:
select id,cast('<M>'+replace([name],',','</M><M>')+'</M>' as xml) as string
from #t
You get the results below:
id string
1 <M>a</M><M>b</M><M>c</M>
2 <M>d</M><M>e</M>
The Node method splits the XML data into relational form. To oversimplify, the Nodes method gives you a table 'Split' with a column 'A'. The Value method finally converts the value in the node to a varchar for projection:
(split.A.value('.','varchar(10)') can be interpreted as Table.Column.Value(root to varchar(10)).
Refer the following links for more:
Nodes method: http://msdn.microsoft.com/en-us/library/ms188282.aspx
Value method: http://msdn.microsoft.com/en-us/library/ms178030.aspx

Pulling a value from Xml in a column

I have a table in Sql server that stores Xml data in one of its columns.
The Xml column data looks like this:
<TestDef Weight="0" FailValue="2" ConceptID="-327">
<ToleranceDef ObjectType="SomeName" TargetValue="0"TargetRange="2" />
</TestDef>
I need to write a query that fetches out all the conceptId's from each rows Xml column.
Here it would be -327
I know I can cast the Xml column to a nvarchar(max) then use some reg exp to get the value but not sure how to use the regular expression
Here's an example using a table variable. It will be the same concept with an actual table:
Declare #XmlTable table (
Id Integer Identity,
XmlValue XML
)
Insert Into #XmlTable (XmlValue) values ('<TestDef Weight="0" FailValue="2" ConceptID="-327"><ToleranceDef ObjectType="SomeName" TargetValue="0" TargetRange="2" /></TestDef>')
Insert Into #XmlTable (XmlValue) values ('<TestDef Weight="0" FailValue="2" ConceptID="-325"><ToleranceDef ObjectType="SomeName" TargetValue="0" TargetRange="2" /></TestDef>')
select
Id,
XmlValue,
XmlValue.value('(/TestDef/#ConceptID)[1]', 'integer') as ConceptId
from
#XmlTable

Update XML Node with Conversion in SQL Server 2005

I have a column that contains XML data, but is TEXT type and not XML type. (I have to leave it like this for another reason).
Basically i need to cast it to NText first and then XML. The only problem is my current format that works for selecting the Node value doesnt work to update it.
Error Message: Incorrect syntax near the keyword 'AS'.
UPDATE tbl_Module_RequestForms_Items
SET CAST(CAST(TicorOregon..tbl_Module_RequestForms_Items.XML AS NTEXT) AS XML).value('(//Record/Submitted)[1]', 'NVARCHAR(max)') = 'True'
WHERE CAST(CAST(TicorOregon..tbl_Module_RequestForms_Items.XML AS NTEXT) AS XML).value('(//Record/Submitted)[1]', 'NVARCHAR(max)') <> 'True'
XML Data:
<Record>
<Submitted>False</Submitted>
</Record>
There might be valid reason to store XML in a [n]varchar(max). If you want to only store the XML it is perfectly OK but if you want to modify parts of the XML using TSQL or if you need to query the XML for values or use node/attribute values in a where clause you should switch to XML where you can benefit from indexes on the data and skip the type conversions. Since text is deprecated you should at least consider to switch data type to [n]varchar(max)
If you had your data in a XML column you would use XML DML to modify the XML. In your case you would use replace value of like this.
update tbl_Module_RequestForms_Items
set XMLData.modify('replace value of (/Record/Submitted/text())[1] with "True"')
where XMLData.value('(/Record/Submitted)[1]', 'bit') = 0
Without the XML data type that is not possible so you have to extract the entire XML document, modify it and then update the table with the modified XML document.
You can of course do that using some kind of client development tool but it is also possible in TSQL.
Declare a table variable with the primary key from tbl_Module_RequestForms_Items and the XMLData column but as data type XML.
Copy the rows from tbl_Module_RequestForms_Items to the table variable that should be updated.
Update the XML using replace value of.
Apply the changes back to tbl_Module_RequestForms_Items.
Something like this where I assume that ID is the primary key for tbl_Module_RequestForms_Items and that your XML data is in column XMLData:
declare #T table
(
ID int primary key,
XMLData xml
)
insert into #T
select M.ID,
M.XMLData
from tbl_Module_RequestForms_Items as M
where cast(cast(XMLData as nvarchar(max)) as xml).value('(/Record/Submitted)[1]', 'bit') = 0
update #T
set XMLData.modify('replace value of (/Record/Submitted/text())[1] with "True"')
update M
set XMLData = cast(T.XMLData as nvarchar(max))
from tbl_Module_RequestForms_Items as M
inner join #T as T
on M.ID = T.ID

SQL Server table to xml

this time i have question how to convert MSSQL table to XML
My source SQL table:
+-----------+-----------------+
|atributname|atributvalue |
+-----------+-----------------+
|phone |222 |
|param4 |bbbbcdsfceecc |
|param3 |bbbbcdsfceecc |
|param2 |bbbbcdsfccc |
+-----------+-----------------+
Expected result sample:
<items>
<phone>222</phone>
<prama4>bbbbcdsfceecc</param4>
<param3>bbbbcdsfceecc</param3>
<param2>bbbbcdsfccc</param2>
</items>
I tried lot of variations of the following query
SELECT atributname,atributvalue
FROM sampletable FOR XML PATH (''), ROOT ('items');
but results are not good :( should be exactly like in "Expected result sample"
any help
ps
Script to create sampletable:
create table sampletable
(atributname varchar(20),
atributvalue varchar(20))
insert into sampletable (atributname,atributvalue)
values ('phone','222');
insert into sampletable (atributname,atributvalue)
values ('param4','bbbbcdsfceecc');
insert into sampletable (atributname,atributvalue)
values ('param3','bbbbcdsfceecc');
insert into sampletable (atributname,atributvalue)
values ('param2','bbbbcdsfccc');
That's not how FOR XML works. It's columns that get turned into XML elements, not rows. In order to obtain the expected result, you would need to have columns named phone, param4, and so on - not rows with these values in attributename.
If there are specific elements you want in the XML, you could perform a pivot on the data first, then use FOR XML.
Example of a pivot would be:
SELECT [phone], [param2], [param3], [param4]
FROM
(
SELECT attributename, attributevalue
FROM attributes
) a
PIVOT
(
MAX(attributevalue)
FOR attributename IN ([phone], [param2], [param3], [param4])
) AS pvt
FOR XML ROOT('items')
Of course the aggregate will only work if attributevalue is a numeric data type. If it's a character-type column, then you'll have some trouble with the pivot, as there are no built-in string aggregates in SQL server AFAIK...
ok
finally i have done this in several ways,
but this is simplest version suitable for medium dataset
declare #item nvarchar(max)
set #item= (SELECT '<' + atributname +'>' +
cast(atributvalue as nvarchar(max)) +'</' + atributname +'>'
FROM sampletable FOR XML PATH (''), ROOT ('items'));
select replace(replace(#item,'<','<'),'>','>')

SQL Server: Output an XML field as tabular data using a stored procedure

I am using a table with an XML data field to store the audit trails of all other tables in the database.
That means the same XML field has various XML information. For example my table has two records with XML data like this:
1st record:
<client>
<name>xyz</name>
<ssn>432-54-4231</ssn>
</client>
2nd record:
<emp>
<name>abc</name>
<sal>5000</sal>
</emp>
These are the two sample formats and just two records. The table actually has many more XML formats in the same field and many records in each format.
Now my problem is that upon query I need these XML formats to be converted into tabular result sets.
What are the options for me? It would be a regular task to query this table and generate reports from it. I want to create a stored procedure to which I can pass that I need to query "<emp>" or "<client>", then my stored procedure should return tabular data.
does this help?
INSERT INTO #t (data) SELECT '
<client>
<name>xyz</name>
<ssn>432-54-4231</ssn>
</client>'
INSERT INTO #t (data) SELECT '
<emp>
<name>abc</name>
<sal>5000</sal>
</emp>'
DECLARE #el VARCHAR(20)
SELECT #el = 'client'
SELECT
x.value('local-name(.)', 'VARCHAR(20)') AS ColumnName,
x.value('.','VARCHAR(20)') AS ColumnValue
FROM #t
CROSS APPLY data.nodes('/*[local-name(.)=sql:variable("#el")]') a (x)
/*
ColumnName ColumnValue
-------------------- --------------------
client xyz432-54-4231
*/
SELECT #el = 'emp'
SELECT
x.value('local-name(.)', 'VARCHAR(20)') AS ColumnName,
x.value('.','VARCHAR(20)') AS ColumnValue
FROM #t
CROSS APPLY data.nodes('/*[local-name(.)=sql:variable("#el")]') a (x)
/*
ColumnName ColumnValue
-------------------- --------------------
emp abc5000
*/
Neither xyz432-54-4231 nor abc5000 is valid XML.
You can try to select only one particular format with a like statement, f.e.:
select *
from YourTable
where YourColumn like '[a-z][a-z][a-z][0-9][0-9][0-9][0-9]'
This would match 3 letters followed by 4 numbers.
A better option is probably to add an extra column to the table, where you save the type of the logging. Then you can use that column to select all "emp" or "client" rows.
An option would be to create a series of views that present the aduit table, per type in the relations that you're execpting
for example
select
c.value('name','nvarchar(50)') as name,
c.value('ssn', 'nvarchar(20)') as ssn
from yourtable
cross apply yourxmlcolumn.nodes('/client') as t(c)
you could then follow the same pattern for the emp
you could also create a view (or computed column) to identify each xml type like this:
select yourxmlcolumn.value('local-name(/*[1])', 'varchar(100)') as objectType
from yourtable
Use open xml method
DECLARE #idoc int
EXEC sp_xml_preparedocument #idoc OUTPUT, #xmldoc
SELECT * into #test
FROM OPENXML (#idoc, 'xmlfilepath',2)
WITH (Name varchar(50),ssn varchar(20)
)
EXEC sp_xml_removedocument #idoc
after you get the data in the #test
and you can manipulate this.
you may be put the diff data in diff xml file.

Resources