On Map Reduce - Part 2

On Map Reduce - Part 1 I introduced the use of a map reduce to retrieve the latest entry on the blog.

The document where we store blog entries looks like

@Document
public class BlogEntry implements Serializable {

    private String id;
    private String subject;
    private String article;
    private Date publishDate;
    private Date lastUpdateDate;
    private String printableHtml;
    //more code continues

}

Spring part as I stated before is pretty straight forward, on my service I added getLastEntry method

@Service
public class BlogEntryService {

    public BlogEntry getLastEntry() {
        MapReduceOptions options = new MapReduceOptions();
        MapReduceResults<MappedBlogEntry> mapReduceResults = mongoTemplate.
                mapReduce("blogEntry", "classpath:mongo/mapallblogs.js" , 
                "classpath:mongo/reducelatestblog.js", MappedBlogEntry.class);
        if (mapReduceResults.iterator().hasNext()) {
            BlogEntry blogEntry = mapReduceResults.iterator().next().getValue();
            return blogEntry;
        } else {
            throw new RuntimeException("No entries for blogEntry");
        }
    }

    @Autowired
    private MongoTemplate mongoTemplate;

}

In order to understand it we need to recall what our map and reduce functions looks like

mapallblog.js >

function () {
    emit("blogEntry", this);
}

on mapallblog it’s emitting the whole object as it is with all it’s properties.

reducelatestblog.js >

function (key, values) {
    var latest = values[0];
    for (var i=0;i<values.length;i++) {
        if (latest.publishDate < values[i].publishDate) {
            latest = values[i];
        }
    }
    return latest;
}

then on reduce function we just select the one with the oldest publishDate and return it.

But what mongo does behind the scenes is map it and pair it with a reduction, hence the result of this called will be an object

{
id:"blogEntry",
value:emittedObject
}

we can map the result with another object

public class MappedBlogEntry {

    private String  id;
    private BlogEntry value;
    //more code continues

}

Going back to the actual spring call

        MapReduceResults<MappedBlogEntry> mapReduceResults = mongoTemplate.
                mapReduce("blogEntry", "classpath:mongo/mapallblogs.js" , 
                "classpath:mongo/reducelatestblog.js", MappedBlogEntry.class);
        if (mapReduceResults.iterator().hasNext()) {
            BlogEntry blogEntry = mapReduceResults.iterator().next().getValue();
            return blogEntry;
        } else {
            throw new RuntimeException("No entries for blogEntry");
        }

mongoTemplate.mapReduce fourth parameter is a class to which map the results this case MappedBlogEntry, and that’s about it. Every time you access home screen a map reduce function is evaluated.