Importing an xml schema into postgres to automatically create a table and then populate it with an xml file?

Question

I have a huge xml file that I cannot open unless I import it into a database. I am using Postgres for this. I have a schema that goes with this data file. There are too many columns so i'd like to automate the process of creating a table from this schema and then importing the data file from the local drive on my computer to populate this table. How do i do this? I saw a lot of answers on SO but haven't been able to understand this correctly. Also I do not have superuser rights so will have to work around that.

Here's what the schema file looks like:

> <?xml version="1.0" encoding="UTF-8"?> <xs:schema
> xmlns:xs="http://www.w3.org/2001/XMLSchema"
> elementFormDefault="qualified"
>     targetNamespace="http://www.drugbank.ca" xmlns="http://www.drugbank.ca">
>     <xs:element name="drugbank" type="drugbank-type">
>         <xs:annotation>
>             <xs:documentation>This is the root element for the DrugBank database schema. DrugBank is a database on drug and
> drug-targets.</xs:documentation>
>         </xs:annotation>
>     </xs:element>
>     <xs:complexType name="drugbank-type">
>         <xs:annotation>
>             <xs:documentation>This is the root element type for the DrugBank database schema.</xs:documentation>
>         </xs:annotation>
>         <xs:sequence>
>             <xs:element name="drug" type="drug-type" maxOccurs="unbounded"/>
>         </xs:sequence>
>         <xs:attribute name="version" type="xs:string" use="required">
>             <xs:annotation>
>                 <xs:documentation>The DrugBank version for the exported XML file.</xs:documentation>
>             </xs:annotation>
>         </xs:attribute>
>         <xs:attribute name="exported-on" type="xs:date" use="required">
>             <xs:annotation>
>                 <xs:documentation>The date the XML file was exported.</xs:documentation>
>             </xs:annotation>
>         </xs:attribute>
>     </xs:complexType>
>     <xs:complexType name="drug-type">
>         <xs:sequence>
>             <xs:element maxOccurs="unbounded" minOccurs="1" name="drugbank-id"
>                 type="drugbank-drug-salt-id-type"> </xs:element>
>             <xs:element name="name" type="xs:string"/>
>             <xs:element name="description" type="xs:string"/>
>             <xs:element name="cas-number" type="xs:string"/>
>             <xs:element name="unii" type="xs:string"/>
>             <xs:element name="average-mass" type="xs:float" minOccurs="0"/>
>             <xs:element name="monoisotopic-mass" type="xs:float" minOccurs="0"/>
>             <xs:element name="state" type="state-type" minOccurs="0"/>
>             <xs:element name="groups" type="group-list-type"/>
>             <xs:element name="general-references" type="reference-list-type"/>
>             <xs:element name="synthesis-reference" type="xs:string"/>
>             <xs:element name="indication" type="xs:string"/>
>             <xs:element name="pharmacodynamics" type="xs:string"/>
>             <xs:element name="mechanism-of-action" type="xs:string"/>
>             <xs:element name="toxicity" type="xs:string"/>
>             <xs:element name="metabolism" type="xs:string"/>
>             <xs:element name="absorption" type="xs:string"/>
>             <xs:element name="half-life" type="xs:string"/>
>             <xs:element name="protein-binding" type="xs:string"/>
>             <xs:element name="route-of-elimination" type="xs:string"/>
>             <xs:element name="volume-of-distribution" type="xs:string"/>
>             <xs:element name="clearance" type="xs:string"/>
>             <xs:element name="classification" type="classification-type" minOccurs="0"/>
>             <xs:element name="salts" type="salt-list-type"/>
>             <xs:element name="synonyms" type="synonym-list-type"/>
>             <xs:element name="products" type="product-list-type"/>
>             <xs:element name="international-brands" type="international-brand-list-type"/>
>             <xs:element name="mixtures" type="mixture-list-type"/>
>             <xs:element name="packagers" type="packager-list-type"/>
>             <xs:element name="manufacturers" type="manufacturer-list-type"/>
>             <xs:element name="prices" type="price-list-type"/>
>             <xs:element name="categories" type="category-list-type"/>
>             <xs:element name="affected-organisms" type="affected-organism-list-type"/>
>             <xs:element name="dosages" type="dosage-list-type"/>
>             <xs:element name="atc-codes" type="atc-code-list-type"/>
>             <xs:element name="ahfs-codes" type="ahfs-code-list-type"/>
>             <xs:element name="pdb-entries" type="pdb-entry-list-type"/>
>             <xs:element name="fda-label" type="xs:anyURI" minOccurs="0"/>
>             <xs:element name="msds" type="xs:anyURI" minOccurs="0"/>
>             <xs:element name="patents" type="patent-list-type"/>
>             <xs:element name="food-interactions" type="food-interaction-list-type"/>
>             <xs:element name="drug-interactions" type="drug-interaction-list-type"/>
>             <xs:element minOccurs="0" name="sequences" type="sequence-list-type"/>
>             <xs:element minOccurs="0" name="calculated-properties" type="calculated-property-list-type"/>
>             <xs:element name="experimental-properties" type="experimental-property-list-type"/>
>             <xs:element name="external-identifiers" type="external-identifier-list-type"/>
>             <xs:element name="external-links" type="external-link-list-type"/>
>             <xs:element name="pathways" type="pathway-list-type"/>
>             <xs:element name="reactions" type="reaction-list-type"/>
>             <xs:element name="snp-effects" type="snp-effect-list-type"/>
>             <xs:element name="snp-adverse-drug-reactions" type="snp-adverse-drug-reaction-list-type"/>
>             <xs:element name="targets" type="target-list-type"/>
>             <xs:element name="enzymes" type="enzyme-list-type"/>
>             <xs:element name="carriers" type="carrier-list-type"/>
>             <xs:element name="transporters" type="transporter-list-type"/>
>         </xs:sequence>

This is only a part of it. It's a huge file. Any kind of help/guidance is much appreciated.

Hey there. How big is this xml file? Coincidently I am now importing a 120GB xml file into my database, but I am using another approach based splitting the xml file, importing them into temporary tables and unnesting them to the target table. Not sure if it is what you want. — Jim Jones, Apr 16 '18 at 16:58
@JimJones Wow! In comparison my file is a meager 725MB.I guess, in my case, I could get away with splitting the file. I've imported csv and text data into the db before but with a lot smaller number of columns. If I'm able to figure out how to import this xml schema into a table, half the battle is won. — serendipity, Apr 16 '18 at 17:51

Jim Jones · Answer 1 · 2018-04-27T17:49:54.977

There are probably a thousand ways to import XML files into PostgreSQL, but here is an alternative I find quite easy to implement and is already tested with large xml documents (120GB+)

Depending on the size of your XML file, consider splitting it. A terrific tool to do so is xml_split. This command splits file.xml in smaller files with a maximum of 100MB:

xml_split -n 5 -l 1 -s 100MB file.xml

Once you have your files split in a reasonable size, you can start importing them without having the risk of running out of memory.

Let's consider the following XML file structure ...

<?xml version="1.0"?>
<t>
    <foo>
        <id j="a">1</id>
        <val>bar1</val>
    </foo>
    <foo>
        <id j="b">8</id>
        <val>bar1</val>
    </foo>
    <foo>
        <id j="c">5</id>
        <val>bar1</val>
    </foo>
    <foo>
        <id j="b">2</id>
    </foo>
</t>

... and the following target table, where we will insert the XML records.

CREATE TABLE t (id TEXT, entry XML);

The code bellow imports XML files into a temporary unlogged table and unnest them into the table t using a CTE (aka WITH clause) by the node <foo>. The command perl -pe 's/\n/\\n/g' replaces newline characters with \\n so that you do not get a Premature end of data exception:

#!/bin/bash

psql testdb -c "CREATE UNLOGGED TABLE tmp (entry xml);"

for f in /path/to/your/files/;do

    cat $f | perl -pe 's/\n/\\n/g' |psql testdb -c "COPY tmp FROM STDIN;"
    psql testdb -c "
    WITH j AS (
      SELECT UNNEST(XPATH('//t/foo',entry)) AS entry FROM tmp
    )
      INSERT INTO t 
      SELECT XPATH('//foo/id/text()',j.entry),j.entry FROM j;

      TRUNCATE TABLE tmp;"

done

psql testdb -c "DROP TABLE tmp;"

And here is your data:

testdb=# SELECT * FROM t;
 id  |          entry           
-----+--------------------------
 {1} | <foo>                   +
     |         <id j="a">1</id>+
     |         <val>bar1</val> +
     |     </foo>
 {8} | <foo>                   +
     |         <id j="b">8</id>+
     |         <val>bar1</val> +
     |     </foo>
 {5} | <foo>                   +
     |         <id j="c">5</id>+
     |         <val>bar1</val> +
     |     </foo>
 {2} | <foo>                   +
     |         <id j="b">2</id>+
     |     </foo>
(4 Zeilen)

Importing an xml schema into postgres to automatically create a table and then populate it with an xml file?

1 Answers1

Linked