Nested XML to joined MySQL Tables - mysql

Question

I have some data with nested fields that I would like to import into MySQL. A lot of files, potentially, so any repeatable scripting language is appreciated. It seems like this should be easier than I am making it, but I can't find a good answer.
I believe the cleanest way would be with joined tables, though it would be nice to have one element ALSO present in the parent table, say if it had the kind code "A1" in the sample below.
A similar query was answered here Parsing nested xml into denormalized table except that wasn't MySQL and that data came with a unique identifier. One of the challenges of my data is that there is no unique identifier in the data to create the primary key for joining tables.
Sample data below. Here, the doc-id and assignor tags would have to be separate tables and joined. The data has a DTD that I'm not including for what it's worth. Any input is much appreciated!
<?xml version="1.0" encoding="UTF-8"?>
<assignment>
<assignment-record>
<reel-no>28879</reel-no>
<frame-no>97</frame-no>
<last-update-date><date>20120903</date></last-update-date>
<recorded-date><date>20120830</date></recorded-date>
<page-count>4</page-count>
<correspondent>
<name>LEE, HONG, DEGERMAN, KANG & WAIMEY</name>
<address-1>660 S. FIGUEROA ST., 23RD FL.</address-1>
<address-2>LOS ANGELES, CA 90017</address-2>
</correspondent>
<conveyance-text>ASSIGNMENT OF ASSIGNORS INTEREST (SEE DOCUMENT FOR DETAILS).</conveyance-text>
</assignment-record>
<assignors>
<assignor>
<name>WOO, SUNGHO</name>
<execution-date><date>20120806</date></execution-date>
</assignor>
<assignor>
<name>CHOI, JAEYOUNG</name>
<execution-date><date>20120806</date></execution-date>
</assignor>
</assignors>
<docproperties>
<property>
<document-id>
<country>US</country>
<doc-number>13277056</doc-number>
<kind>X0</kind>
<date>20111019</date>
</document-id>
<document-id>
<country>US</country>
<doc-number>20120213136</doc-number>
<kind>A1</kind>
<date>20120823</date>
</document-id>
<title lang="en">SYSTEMS AND METHODS FOR CONTROLLING SENSOR DEVICES IN MOBILE DEVICES</title>
</property>
</docproperties>
</assignment>
</xml>

Answer 1

Since this three-year old unanswered question was recently bumped up by another user, I went ahead and answered it on behalf of the community as the original OP is no longer active.
For future readers, anytime you need to flatten nested XML files for flatfile import such as database tables consider XSLT, the transformation language used to manipulate XML files. Practically all general purpose languages has some library for XSLT 1.0 processing including Python, PHP, Perl, Java, C#, VB, and others.
As OP mentioned Python, below examples uses the third-party lxml module to flatten the XML file. And to generate a unique id to relate the <assignor> and <document-id> data, an XPath in XSLT script is run using ancestor to retrieve the the reel-no value from <assignment-record> which both nodes share as a sibling. This is similar to the TSQL solution in posted link.
XSLT Scripts (save as .xsl to be referenced in Python)
Assignor Transformation
<xsl:transform xmlns:xsl="http://www.w3.org/1999/XSL/Transform" version="1.0">
<xsl:output version="1.0" encoding="UTF-8" indent="yes" />
<xsl:strip-space elements="*"/>
<xsl:template match="/assignment">
<xsl:copy>
<xsl:apply-templates select="descendant::assignor"/>
</xsl:copy>
</xsl:template>
<xsl:template match="assignor">
<xsl:copy>
<assign_id>
<xsl:value-of select="ancestor::assignment/assignment-record/reel-no"/>
</assign_id>
<xsl:copy-of select="name"/>
<xsl:copy-of select="execution-date/date"/>
</xsl:copy>
</xsl:template>
</xsl:transform>
Document-id Transformation
<xsl:transform xmlns:xsl="http://www.w3.org/1999/XSL/Transform" version="1.0">
<xsl:output version="1.0" encoding="UTF-8" indent="yes" />
<xsl:strip-space elements="*"/>
<xsl:template match="/assignment">
<xsl:copy>
<xsl:apply-templates select="descendant::document-id"/>
</xsl:copy>
</xsl:template>
<xsl:template match="document-id">
<xsl:copy>
<assign_id>
<xsl:value-of select="ancestor::assignment/assignment-record/reel-no"/>
</assign_id>
<xsl:copy-of select="*"/>
</xsl:copy>
</xsl:template>
</xsl:transform>
Python Script (transforms source into two files)
import lxml.etree as ET
dom = ET.parse('Assignment.xml')
xslt = ET.parse('Assignor_XSLT_Script.xsl')
transform = ET.XSLT(xslt)
newdom = transform(dom)
xmlfile = open(os.path.join(cd, 'Assignor.xml'),'wb')
xmlfile.write(newdom)
xmlfile.close()
xslt = ET.parse('Document-Id_XSLT_Script.xsl')
transform = ET.XSLT(xslt)
newdom = transform(dom)
xmlfile = open(os.path.join(cd, 'Document-Id.xml'),'wb')
xmlfile.write(newdom)
xmlfile.close()
XML Outputs (can now use MySQL's LOAD XML effectively)
<?xml version="1.0" encoding="UTF-8"?>
<assignment>
<assignor>
<assign_id>28879</assign_id>
<name>WOO, SUNGHO</name>
<date>20120806</date>
</assignor>
<assignor>
<assign_id>28879</assign_id>
<name>CHOI, JAEYOUNG</name>
<date>20120806</date>
</assignor>
</assignment>
<?xml version="1.0" encoding="UTF-8"?>
<assignment>
<document-id>
<assign_id>28879</assign_id>
<country>US</country>
<doc-number>13277056</doc-number>
<kind>X0</kind>
<date>20111019</date>
</document-id>
<document-id>
<assign_id>28879</assign_id>
<country>US</country>
<doc-number>20120213136</doc-number>
<kind>A1</kind>
<date>20120823</date>
</document-id>
</assignment>