How To Import Excel Data into VuFind

by Jonathan Jacobsen Tuesday, January 08, 2019 4:52 PM

Recently we had a new client come to us looking for help with several subscription-based VuFind sites they manage, and ultimately to have us host them as part of our managed hosting service. This client had a unique challenge for us: 3 million records, available as tab-separated text files of up to 70,000 records each.

Most of the data sets we work with are relatively small: libraries with a few thousand records, archives with a few tens of thousands, and every so often, databases of a few hundred thousand, like those in the Arctic Health bibliography.

While VuFind and the Apache Solr search engine that powers it (and also powers our Andornot Discovery Interface) have no trouble with that volume of records, transforming the data from hundreds of tab-separated text files into something Solr can use, in an efficient manner, was a pleasant challenge.

VuFind has excellent tools for importing traditional library MARC records, using the SolrMarc tool to post data to Solr. For other types data, such as records exported from DB/TextWorks databases, we’ve long used the PHP-based tools in VuFind that use XSLTs to transform XML into Solr's schema and post it to Solr. While this has worked well, XSLTs are especially difficult to debug, so we considered alternatives.

For this new project, we knew we needed to write some code to manipulate the 3 million records in tab-separated text files into XML, and we knew from our extensive experience with Solr that it's best to post small batches of records at a time, in separate files, rather than one large post of 3 million! So we wrote a python script to split up the source data into separate files of about 1,000 records each, and also remove invalid characters that had crept in to the data over time (this data set goes back decades and has likely been stored in many different character encodings on many different systems, so it's no surprise there were some gremlins).

Once the script was happily creating Solr-ready XML files, rather than use VuFind's PHP tools and an XSLT to index the data, it just seemed more straightforward to push the XML directly to Solr. For this, we wrote a bash shell script that uses the post tool that ships with Solr to iterate through the thousands of data files and push each to Solr, logging the results.

The combination of a python script to convert the tab-separated text files into Solr-ready XML and a bash script to push it to Solr worked extremely well for this project. Python is lightning fast at processing text and pushing data directly to Solr is definitely faster than invoking XSLT transformations.

This approach would work well for any data. Python is a very forgiving language to develop with, making it easy and quick to write scripts to process any data source. In fact, since this project, we've used Python to manipulate a FileMaker Pro database export for indexing in our Andornot Discovery Interface (also powered by Apache Solr) and to harvest data from the Internet Archive and Online Archive of California, for another Andornot Discovery Interface project (watch this blog for news of both when they launch).

We look forward to more challenges like this one! Contact us for help with your own VuFind, Solr and similar projects.

Java 11 date parsing? Locale, locale, locale.

by Peter Tyrrell Monday, January 07, 2019 11:39 AM

Java is undergoing some considerable licensing changes, prompting us to plan an all-out move from Oracle Java 8 to OpenJDK Java 11 this Spring for every Solr instance we host. I have been running covertly about the hills setting traps for Java 11.0.1 to see what I might snare before unleashing it on our live servers. I caught something this week.

Dates! Of course it's about parsing dates! I noticed that the Solr Data Import Handler (DIH) transforms didn't handle making created dates during ingest. (In DIH, we use a script transformer and manipulate some Java classes with javascript. This includes the parsing of dates from text.) Up until now, our DIH has used an older method of parsing dates with a Java class called SimpleDateFormat. If you look for info on parsing dates in Java, you will find years and years of advice related to that class and its foibles, and then you will notice that in recent times experts advise using the java.time classes introduced in Java 8. Since SimpleDateFormat didn't work during DIH, I assumed that SimpleDateFormat was deprecated in Java 11 (it isn't actually), and moved to convert the relevant DIH code to use java.time.

Many hours passed here, during which the output of two lines of code* made no goddamn sense at all. The javadocs that describe the behaviour of java.time classes are completely inadequate, with their stupid little "hello, world" examples, when dates are tricky, slippery, malicious dagger-worms of pure hatred. Long story short, a date like '2004-09-15 12:00:00 AM' produced by Inmagic ODBC from a DB/Textworks database could not be parsed. The parser choked on the string at "AM," even though my match pattern was correct: 'uuuu-MM-dd hh:mm:ss a'. Desperate to find the tiniest crack to exploit, I changed every variable I could think of one at a time. That was how I found that, when I switched to Java 8, the same exact code worked. Switch back to Java 11. Not working. Back to Java 8. Working. WTF?

I thought, maybe the Nashorn scripting engine that allows javascript to be interpreted inside the Java JVM is to blame, because this scenario does involve Java inside javascript inside Java, which is weird. So I set up a Java project with Visual Studio Code and Maven and wrote some unit tests in pure Java. (That was pretty fun. It was about the same effort as ordering a pizza in Italian when you don’t speak Italian: everything about the ordering process was tantalizingly familiar but different enough to delay my pizza for quite some time.) The problem remained: parsing worked as-written in Java 8, but not Java 11.

I started writing a Stack Overflow question. In so doing, I realized I hadn't tried an overload method of java.time.format.DateTimeFormatter.ofPattern() which takes a locale. I had already dotted many i's and crossed a thousand t's, but I wanted to really impress anyone reading the question that I had done my homework, because I hate looking ignorant, so I wrote another unit test that passed in Locale.ENGLISH and, ohmigawd, that solved the problem entirely. If you have been following along, that means that "AM/PM" could not be understood by the parser, even with the right pattern matcher, without the context of a locale, and obviously the default locale used by the simpler version of DateTimeFormatter.ofPattern() was inadequate to the task. I tested and Locale.ENGLISH and Locale.US both worked with "AM/PM" but Locale.CANADA did not. Likely the latter is my default locale, because I do reside in Canada. Really? Really, Java? We have AM and PM here in the Great White North, I assure you.

I don’t know if this a bug in Java 11. I’m merely happy to have understood the problem at this point. Just another day in the developer life, eh? Something that should be a snap becomes a grueling carnival ride that deposits you at the exit, white-faced and shaking, with an underwhelming sense of minor accomplishment. How do you explain to people that you spent 8 hours teaching a computer to treat an ordinary date as a date? Write a blog post, I guess. Winking smile

* Two lines of code. 8 hours of frustration. Here it is, ready?

import java.time.LocalDateTime;
import java.time.format.DateTimeFormatter;
import java.util.Locale;

public class App {

    public LocalDateTime Parse (String dateText, String pattern) {

        DateTimeFormatter parser = DateTimeFormatter.ofPattern(pattern, Locale.ENGLISH);
        LocalDateTime date = LocalDateTime.parse(dateText, parser);
        return date;

    }
}

Month List