SQL Select FOR XML into Solr document
I'm trying to get a SQL select statement to generate XML which conforms to the Solr standard.
Given a table like:
id | name
---------
1 | one
2 | two
3 | three
I need a result which is like (with or without the root node):
<add>
<doc>
<field name="id">1</field>
<field name="name">one</field>
<开发者_C百科;/doc>
<doc>
<field name="id">2</field>
<field name="name">two</field>
</doc>
<doc>
<field name="id">3</field>
<field name="name">three</field>
</doc>
</add>
Is it possible to generate that structure using a FOR XML
query, or will I need an XSLT or some other mechanism to match that schema?
Here's a slightly different way using constructors.
DECLARE @sample TABLE
(
[id] int NOT NULL,
[name] varchar(50) NOT NULL
);
INSERT INTO @sample ([id], [name])
SELECT 1, 'one' UNION ALL
SELECT 2, 'two' UNION ALL
SELECT 3, 'three';
SELECT
CONVERT(xml, N'').query
(N'
<doc>
{
element field
{
attribute name {"id"},
text{sql:column("id")}
},
element field
{
attribute name {"name"},
text{sql:column("name")}
}
}
</doc>
')
FROM
@sample
FOR XML PATH(N''), ROOT(N'add');
* EDIT: Just thought of another way to do this (but still requires knowledge of the columns ahead of time) * Again, i'm uncertain of the performance implications of either approach.
SELECT
(
SELECT
'id' AS [@name],
[id] AS [data()]
FOR XML PATH('field'), TYPE
) AS [*],
(
SELECT
'name' AS [@name],
[name] AS [data()]
FOR XML PATH('field'), TYPE
) AS [*]
FROM
@sample
FOR XML PATH(N'doc'), ROOT(N'add');
* UPDATE 2: Dynamic yet non-performant method inspired by Aaron Bertrand's comments *
This was a proof of concept of the method described in the post Aaron referenced in his comments. (It performs horribly on larger datasets)
-- Inspired by Aaron Bertrand's comment
WITH [cte_KVP]
AS
(
-- Generating Key/Value pairs for columns in a table
-- Courtesey of Mikael Eriksson (http://stackoverflow.com/questions/7341143/flattening-of-a-1-row-table-into-a-key-value-pair-table/)
SELECT
[T2].[N].value(N'local-name(.)', N'sysname') AS [Key],
[T2].[N].value(N'.', N'nvarchar(max)') AS [Value],
[T2].[N].value(N'../GROUP[1]', N'int') AS [GROUP] -- 3. Used for to group the key/value pairs per row
FROM
(
SELECT
*,
ROW_NUMBER() OVER (ORDER BY (SELECT 0)) AS [GROUP] -- 1. Generating a simple "identity" value.
FROM
@sample
FOR XML PATH(N'Row'), TYPE -- 2. Adding the 'Row' to the path separates each row, and allows us to backtrack via xpath to get the "GROUP" id
) AS [T1]([x])
CROSS APPLY
[T1].[x].nodes(N'Row/*') AS [T2]([N])
WHERE
[T2].[N].value(N'local-name(.)', N'sysname') <> N'GROUP'
)
SELECT
[InnerNodes].[xml] AS [*]
FROM
(
-- Probably preferable to use a table of numbers here
SELECT DISTINCT
[GROUP]
FROM
[cte_KVP]
) AS [Numbers]([Number])
CROSS APPLY
(
-- Generating the xml fragment specified by OP
SELECT
[cte_KVP].[Key] AS [@name],
[cte_KVP].[Value] AS [data()]
FROM
[cte_KVP]
WHERE
[cte_KVP].[GROUP] = [Numbers].[Number]
FOR XML PATH(N'field'), ROOT(N'doc'), TYPE
) AS [InnerNodes]([xml])
FOR XML PATH(N''), ROOT(N'add');
It's probably not as "natural" as you want, and if you don't know the column names in advance, you'd have to build it dynamically, but this seems to produce the doc you're after:
SELECT
CONVERT(XML, '<field name="id">' + RTRIM(id) + '</field>'
+ '<field name="name">' + name + '</field>')
FROM dbo.[table]
FOR XML PATH(N'doc'), ROOT(N'add');
And here is a dynamic approach:
DECLARE
@table NVARCHAR(512) = N'dbo.[table]',
@sql NVARCHAR(MAX) = N'';
SELECT @sql += '
+ ''<field name="' + name
+ '">'' + CONVERT(NVARCHAR(MAX), '
+ QUOTENAME(name) + ') + ''</field>'''
FROM sys.columns
WHERE object_id = OBJECT_ID(@table);
SET @sql = 'SELECT CONVERT(XML, ' + STUFF(@sql, 1, 4, '')
+ ') FROM ' + @table
+ ' FOR XML PATH(N''doc''), ROOT(N''add'');';
PRINT @sql;
-- EXEC sp_executesql @sql;
I am assuming that you are programming in .NET since you are using SQL Server. If you are have you considered using the SolrNet client to load documents into your Solr server?
精彩评论