如何将HTML包装的字符串从SQL表中的列提取到新表中？
我在SQL Server 2005表中有一个名为BIO的列—BIO列中的数据格式如下：如何将HTML包装的字符串从SQL表中的列提取到新表中？,sql,sql-server-2005,stored-procedures,complexity-theory,substring,Sql,Sql Server 2005,Stored Procedures,Complexity Theory,Substring,我在SQL Server 2005表中有一个名为BIO的列—BIO列中的数据格式如下： <HTML><HEAD><TITLE></TITLE></HEAD><BODY><STRONG><A name=SN>AARTS</A>, <A name=GN>Michelle Marie</A>, </STRONG><A name=HO>B.Sc.&l
<HTML><HEAD><TITLE></TITLE></HEAD><BODY><STRONG><A name=SN>AARTS</A>, <A name=GN>Michelle Marie</A>, </STRONG><A name=HO>B.Sc.</A>, <A name=HO>M.Sc.</A>, <A name=HO>Ph.D.</A>; <A name=OC>scientist, professor</A>; b. <A name=BC>St. Marys</A>, Ont. <A name=BY>1970</A>; <A name=PA>d. Wm. and H. Aarts</A>; <A name=ED>e. Univ. of Western Ont. B.Sc.(Hons.) 1994, M.Sc. 1997</A>; <A name=ED>McGill Univ. Ph.D. 2002</A>; <A name=MA>m. L. MacManus</A>; two children; <A name=PO>CANADA RESEARCH CHAIR IN SIGNAL TRANSDUCTION IN ISCHEMIA</A> and <A name=PO>ASST. PROF., DEPT. OF BIOL. SCI., UNIV. OF TORONTO SCARBOROUGH 2006– </A>; Postdoctoral Fellow, Toronto Western Hosp. 2000–06; Expert Cons., Auris Med. SAS, Montpellier, France; mem., Centre for the Neurobiol. of Stress; named INMHA Brainstar of the Year 2003; Bd. of Dirs. & Fundraising Chair, N'Sheemaehn Childcare; mem., Soc. for Neurosci.; Cdn. Physiol. Soc.; Cdn. Assn. for Neurosci.; <A name=WK>co-author: 'Therapeutic Tools in Brain Damage' in <EM>Proteomics and Protein Interactions: Biology, Chemistry, Bioinformatics and Drug Design </EM>2005; 18 pub. journal articles</A>; Office: <A name=OF1_L1>1265 Military Trail</A>, <A name=OF1_CT>Scarborough</A>, <A name=OF1_PR>Ont.</A> <A name=OF1_PC>M1C 1A4</A>. </BODY></HTML>
CONTACT_ID SN GN HO OC PO DB PA BY ED
3 AARON Raymond Leonard B.Sc. business coach, professional speaker, real estate entrepreneur D>AARON
5 AATAMI Pita C.Q. business executive; Kuujjuaq
7 ABBOTT Anthony C. P.C. lawyer Montreal
8 ABBOTT Elizabeth M.A. historian Ottawa
9 ABBOTT (Caroline) Louise D>ABBOTT writer, photographer, filmmaker Montreal
SELECT CONTACT_ID
,'SN' =
CASE
WHEN CHARINDEX('<A name=SN>', [BIO]) = 0 THEN NULL
ELSE dbo.udf_StripHTML(SUBSTRING([BIO], (CHARINDEX('<A name=SN>', [BIO]) + 11), (CHARINDEX('</A>', [BIO], CHARINDEX('<A name=SN>', [BIO])) - CHARINDEX('<A name=SN>', [BIO])-11)))
END
,'GN' =
CASE
WHEN CHARINDEX('<A name=GN>', [BIO]) = 0 THEN NULL
ELSE dbo.udf_StripHTML(SUBSTRING([BIO], (CHARINDEX('<A name=GN>', [BIO]) + 11), (CHARINDEX('</A>', [BIO], CHARINDEX('<A name=GN>', [BIO])) - CHARINDEX('<A name=GN>', [BIO])-11)))
END
,'HO' =
CASE
WHEN CHARINDEX('<A name=HO>', [BIO]) = 0 THEN NULL
ELSE dbo.udf_StripHTML(SUBSTRING([BIO], (CHARINDEX('<A name=HO>', [BIO]) + 11), (CHARINDEX('</A>', [BIO], CHARINDEX('<A name=HO>', [BIO])) - CHARINDEX('<A name=HO>', [BIO])-11)))
END
,'OC' =
CASE
WHEN CHARINDEX('<A name=OC>', [BIO]) = 0 THEN NULL
ELSE dbo.udf_StripHTML(SUBSTRING([BIO], (CHARINDEX('<A name=OC>', [BIO]) + 11), (CHARINDEX('</A>', [BIO], CHARINDEX('<A name=OC>', [BIO])) - CHARINDEX('<A name=OC>', [BIO])-11)))
END
,'PO' =
CASE
WHEN CHARINDEX('<A name=PO>', [BIO]) = 0 THEN NULL
ELSE dbo.udf_StripHTML(SUBSTRING([BIO], (CHARINDEX('<A name=PO>', [BIO]) + 11), (CHARINDEX('</A>', [BIO], CHARINDEX('<A name=PO>', [BIO])) - CHARINDEX('<A name=PO>', [BIO])-11)))
END
,'BD' =
CASE
WHEN CHARINDEX('<A name=BD>', [BIO]) = 0 THEN NULL
ELSE dbo.udf_StripHTML(SUBSTRING([BIO], (CHARINDEX('<A name=BD>', [BIO]) + 11), (CHARINDEX('</A>', [BIO], CHARINDEX('<A name=BD>', [BIO])) - CHARINDEX('<A name=BD>', [BIO])-11)))
END
,'PA' =
CASE
WHEN CHARINDEX('<A name=PA>', [BIO]) = 0 THEN NULL
ELSE dbo.udf_StripHTML(SUBSTRING([BIO], (CHARINDEX('<A name=PA>', [BIO]) + 11), (CHARINDEX('</A>', [BIO], CHARINDEX('<A name=PA>', [BIO])) - CHARINDEX('<A name=PA>', [BIO])-11)))
END
,'BY' =
CASE
WHEN CHARINDEX('<A name=BY>', [BIO]) = 0 THEN NULL
ELSE dbo.udf_StripHTML(SUBSTRING([BIO], (CHARINDEX('<A name=BY>', [BIO]) + 11), (CHARINDEX('</A>', [BIO], CHARINDEX('<A name=BY>', [BIO])) - CHARINDEX('<A name=BY>', [BIO])-11)))
END
,'ED' =
CASE
WHEN CHARINDEX('<A name=ED>', [BIO]) = 0 THEN NULL
ELSE dbo.udf_StripHTML(SUBSTRING([BIO], (CHARINDEX('<A name=ED>', [BIO]) + 11), (CHARINDEX('</A>', [BIO], CHARINDEX('<A name=ED>', [BIO])) - CHARINDEX('<A name=ED>', [BIO])-11)))
END
--INTO [cww].[dbo].[BioDetails]
FROM [cww].[dbo].[Contacts]
ORDER BY CONTACT_ID
我将创建一个包含联系人id、锚类型和值的表，解析数据并向其中添加记录，然后使用交叉表将其吐出。这将允许您轻松找到所有锚类型，允许多个/无锚类型，等等
您可能需要对数据进行多次传递，或者考虑使用不是主要基于SQL设置的工具，如C.< /P> < P>我不知道如何在纯T-SQL中做到这一点。p> 如果可以在应用程序中检索CONTACT_ID和BIO列，那么可以迭代结果集，将BIO数据解析为XML，然后使用XPath获取name属性值和锚主体，构建要插入到新表中的数据映射。由于您不知道可能存在的所有不同名称，因此每次运行表时可能需要重新创建该表，因此存储在集合中找到的名称，并在遍历所有行后使用集合生成CREATETABLE语句
DB代码是纯粹的幻想，但这里有一个片段，展示了如何使用XOM XML库实现它。我不确定这是否可行，因为您的属性值没有被引用，但您可能会找到一个不太挑剔的解析器，我相信您可以在.NET中执行类似的操作ResultSet results = db.query("select CONTACT_ID, BIO from [cww].[dbo].[Contacts]");
Set<String> newTableColumns = new Set<String>();
newTableColumns.put("CONTACT_ID");
List<Map<String,String> > dataToInsert = new ArrayList<Map<String,String> >();
Builder parser = new Builder();
for (ResultRow resultRow : results) { // iterate over the result set
Map<String,String> rowDataToInsert = new HashMap<String,String>();
rowData.put("CONTACT_ID", resultRow.get("CONTACT_ID"));
// parse the BIO data as an XML document
Document doc = parser.build(resultRow.get("BIO"), "");
// query the document using XPath
Nodes namedAnchors = doc.query("//a[@name]");
for (int nItr = 0; nItr < namedAnchors.size(); nItr++) {
Element anchor = (Element) namedAnchors.get(nItr);
String name = anchor.getAttributeValue("name");
String anchorBody = anchor.getValue();
newTableColumns.put(name);
rowDataToInsert.put(name, anchorBody);
}
// we've stored all the anchor data from this row, so put it away
dataToInsert.add(rowDataToInsert);
}
// create your table
db.createTable("NEW_TABLE_NAME", newTableColumns);
// insert into your new table
db.batchInsert("NEW_TABLE_NAME", dataToInsert);
哎哟我认为这已经成为我所见过的最好的建议，因为它不在数据库中存储HTML，而是希望将数据取回。如前所述，此表还存在大量其他问题-我建议创建一组表来永久保存要提取的数据，根据需要创建手动查询，是的，您必须在执行时进行验证。。。。当你全部完成后，删除那个东西…是的，与别人的数据一起工作的缺点。这项工作的全部目的是将BIO列转换为关系数据，并使用可重用脚本为每种类型的锚点提供查找表，因为我们每年都会得到一份新的数据副本，这些数据需要进行关系化处理，以便更好地在网站上搜索……难道sql server的家伙们没有字段类型的XML吗？用它来存储XML有趣的是，我没有意识到这一点，因为我不是一个“sql server”的家伙。。。我会调查一下，看看这对我是否有帮助。非常感谢。另外，这是我获取数据的方式，这不是我的原因，感谢您的输入-尤其是C可能比我高出很多，您是否有可能提供一个代码示例，用于将锚名称解析为锚列？以及你所说的“交叉表”。我不是一个真正的SQL专家，我更喜欢html、css、js、php、.net、ruby等。。但是有时候工作中会有一些SQL…非常感谢。我想我要试试这样的东西。我会把结果告诉你。
SELECT CONTACT_ID
,'SN' =
CASE
WHEN CHARINDEX('<A name=SN>', [BIO]) = 0 THEN NULL
ELSE dbo.udf_StripHTML(SUBSTRING([BIO], (CHARINDEX('<A name=SN>', [BIO]) + 11), (CHARINDEX('</A>', [BIO], CHARINDEX('<A name=SN>', [BIO])) - CHARINDEX('<A name=SN>', [BIO])-11)))
END
,'GN' =
CASE
WHEN CHARINDEX('<A name=GN>', [BIO]) = 0 THEN NULL
ELSE dbo.udf_StripHTML(SUBSTRING([BIO], (CHARINDEX('<A name=GN>', [BIO]) + 11), (CHARINDEX('</A>', [BIO], CHARINDEX('<A name=GN>', [BIO])) - CHARINDEX('<A name=GN>', [BIO])-11)))
END
,'HO' =
CASE
WHEN CHARINDEX('<A name=HO>', [BIO]) = 0 THEN NULL
ELSE dbo.udf_StripHTML(SUBSTRING([BIO], (CHARINDEX('<A name=HO>', [BIO]) + 11), (CHARINDEX('</A>', [BIO], CHARINDEX('<A name=HO>', [BIO])) - CHARINDEX('<A name=HO>', [BIO])-11)))
END
,'OC' =
CASE
WHEN CHARINDEX('<A name=OC>', [BIO]) = 0 THEN NULL
ELSE dbo.udf_StripHTML(SUBSTRING([BIO], (CHARINDEX('<A name=OC>', [BIO]) + 11), (CHARINDEX('</A>', [BIO], CHARINDEX('<A name=OC>', [BIO])) - CHARINDEX('<A name=OC>', [BIO])-11)))
END
,'PO' =
CASE
WHEN CHARINDEX('<A name=PO>', [BIO]) = 0 THEN NULL
ELSE dbo.udf_StripHTML(SUBSTRING([BIO], (CHARINDEX('<A name=PO>', [BIO]) + 11), (CHARINDEX('</A>', [BIO], CHARINDEX('<A name=PO>', [BIO])) - CHARINDEX('<A name=PO>', [BIO])-11)))
END
,'BD' =
CASE
WHEN CHARINDEX('<A name=BD>', [BIO]) = 0 THEN NULL
ELSE dbo.udf_StripHTML(SUBSTRING([BIO], (CHARINDEX('<A name=BD>', [BIO]) + 11), (CHARINDEX('</A>', [BIO], CHARINDEX('<A name=BD>', [BIO])) - CHARINDEX('<A name=BD>', [BIO])-11)))
END
,'PA' =
CASE
WHEN CHARINDEX('<A name=PA>', [BIO]) = 0 THEN NULL
ELSE dbo.udf_StripHTML(SUBSTRING([BIO], (CHARINDEX('<A name=PA>', [BIO]) + 11), (CHARINDEX('</A>', [BIO], CHARINDEX('<A name=PA>', [BIO])) - CHARINDEX('<A name=PA>', [BIO])-11)))
END
,'BY' =
CASE
WHEN CHARINDEX('<A name=BY>', [BIO]) = 0 THEN NULL
ELSE dbo.udf_StripHTML(SUBSTRING([BIO], (CHARINDEX('<A name=BY>', [BIO]) + 11), (CHARINDEX('</A>', [BIO], CHARINDEX('<A name=BY>', [BIO])) - CHARINDEX('<A name=BY>', [BIO])-11)))
END
,'ED' =
CASE
WHEN CHARINDEX('<A name=ED>', [BIO]) = 0 THEN NULL
ELSE dbo.udf_StripHTML(SUBSTRING([BIO], (CHARINDEX('<A name=ED>', [BIO]) + 11), (CHARINDEX('</A>', [BIO], CHARINDEX('<A name=ED>', [BIO])) - CHARINDEX('<A name=ED>', [BIO])-11)))
END
--INTO [cww].[dbo].[BioDetails]
FROM [cww].[dbo].[Contacts]
ORDER BY CONTACT_ID
ResultSet results = db.query("select CONTACT_ID, BIO from [cww].[dbo].[Contacts]");
Set<String> newTableColumns = new Set<String>();
newTableColumns.put("CONTACT_ID");
List<Map<String,String> > dataToInsert = new ArrayList<Map<String,String> >();
Builder parser = new Builder();
for (ResultRow resultRow : results) { // iterate over the result set
Map<String,String> rowDataToInsert = new HashMap<String,String>();
rowData.put("CONTACT_ID", resultRow.get("CONTACT_ID"));
// parse the BIO data as an XML document
Document doc = parser.build(resultRow.get("BIO"), "");
// query the document using XPath
Nodes namedAnchors = doc.query("//a[@name]");
for (int nItr = 0; nItr < namedAnchors.size(); nItr++) {
Element anchor = (Element) namedAnchors.get(nItr);
String name = anchor.getAttributeValue("name");
String anchorBody = anchor.getValue();
newTableColumns.put(name);
rowDataToInsert.put(name, anchorBody);
}
// we've stored all the anchor data from this row, so put it away
dataToInsert.add(rowDataToInsert);
}
// create your table
db.createTable("NEW_TABLE_NAME", newTableColumns);
// insert into your new table
db.batchInsert("NEW_TABLE_NAME", dataToInsert);