PostgreSQL，Epidoc和Dublin核心-数据库和xml

Question

I'm about to start a project which records textual information using Epidoc xml. Here is an example: http://www.stoa.org/epidoc/gl/latest/supp-structure.html I'm wanting to store the data in Postgresql. I understand xml and I understand the basics of postgresql. What is the correct/best way to put these two things together?

For example, so I can use sql to select * from db where xmltag = value

Answer 1

Very short and simplified mini-primer

Create your tables, they will look something like

CREATE TABLE xml_table 
(
    document_id integer /* you'd normally use serial */ PRIMARY KEY,
    xml_data xml
) ;

Check PostgreSQL documentation about the XML data type .

You will fill your tables with queries like the following ones:

/* If you use XML as content, you'd insert it this way */
INSERT INTO
    xml_table (document_id, xml_data)
VALUES
        (1, xmlparse(content '<doc><title>Doc title</title></doc>')),

    (2, xmlparse(content '<doc>
          <preface>This is the preface</preface>
             <chapter><title>Hello</title><content>This is a content</content></chapter>
         <chapter><title>Good Bye</title><content>This is the end</content></chapter>
     </doc>')),
    (3, xmlparse(content '<doc>
         <preface>Yet a preface</preface>
             <chapter><title>C1</title><content>Content of C1</content></chapter>
         <chapter><title>C2</title><content>Content of C2</content></chapter>
     </doc>')) ;

I am not using EpiDocs as examples at this point for the sake of conciseness, but the concept is the same.

Note that, normally, you do not want to store your whole database as a single XML document (that would be inefficient for most DBs), but as several documents, identified by a number (or whatever is more convenient to use to identify them)

If you insert whole documents (and EpiDoc seems to require this approach):

/* If your XML are documents, this way */
INSERT INTO
    xml_table (document_id, xml_data)
VALUES
    (4, xmlparse(document '<?xml version="1.0"?><book><title>Oh my God</title><content>Short book</content></book>')) ;

Note that PostgreSQL will not check that your document complies with your DTD (this would require the database to query the outside world , which is normally out of the scope of databases). You have to check for conformance, if needed, in your software before inserting the values to the DB, if you want to ensure.

You'll retrieve whole document (or content) this way:

SELECT
   xml_data
FROM
   xml_table
WHERE
    document_id = 3 ;

Although you'll normally query using xpath and xpath_exists to get specific items. For instance, imagine you want to get the title of the last chapter of each book (having chapters). You'd use:

SELECT
    /* Get the text content of title of the last chapter of every doc */
    xpath('/doc/chapter[last()]/title/text()', xml_data) AS result
FROM
    xml_table 
WHERE
    /* Choose only the docs where they have (at least) a chapter with title */
    xpath_exists('/doc/chapter/title', xml_data) ;

Check PostgreSQL XML functions, and XPath Intro .