blogs

Now you can create an index via customized fetcher, and select the SalesForce.com adapter, fill in username/password and the object you want to search on, like "Account".

The rest is exactly the same as ordinary database search configuration.

The good thing is, the incremental indexing will also work. You just schedule the job. And on SalesForce.com side, you need to make the object replicable. This replicable setting doesn't matter if you create index from scratch.

As your data grows, one single box can not hold all of your data. So what to do? Buy a bigger box or add another box?

Getting a bigger box may work for some. Actually it's much simpler and cost less on man-hours. A few thousands of dollars are much cheaper than putting several developers on it.

Of course, Sharded Search can go further by just adding more computer boxes.

DBSight can now support sharded search now! And you don't need to do much. Just adjust the search URL!

Suppose you are 2 indexes on 2 different boxes. The URL format is pretty simple, just say:

 ?indexName=index1,index2@hostname2:port2/dbsight

This is an interesting feature!

Usually if you have several DBSight servers, you may have to log on to several servers to copy index configurations. And sometimes it is hard to track which one has the latest change.

For Enterprise licenses, to use the Remote Index Replication feature, you also need to setup where the master index is.

Now everything is just one click away!

If you have the remote DBSight server's password, you can browse the local network, copy the remote index configuration and templates with one click!

One user, Rajendra, asked for a new feature: copy the index configuration.

It is a basic operation that we have missed all along!

DBSight can create, delete, and even upload index configurations. But "copy".... duh! Thanks Rajendra!

We have spent a little time to add it in. Hopefully you will like it. Now you could easily manage all the index configurations, make it easier for you to develop, test, and put DBSight to production.

Please let us know any feature you want. We will add it if it's a good feature.

Brand New feature not seen anywhere else!

Normally when you do a facet search, you get how many hits for that facet. This is a great feature. But there is more!

Now wiht 2.1.6 beta, you can also get Sum and Average for a different field!

It's just like SQL:

  select count(*), sum(f2),avg(f2) 
  from t 
  group by f1

If you are upgrading to 2.1.3, you may need to re-create the whole index. This is because the storage format of time is changed, to allow support for time before 1970.

This is a rather big change, so please pay attention to it.

We have just implemented a scaffold of Google Map!

If you have a table with latitude and longitude, you are just minutes away from a working search with Google Map.

Well, it's only in beta, because there are some problems with IE. If you find out what's wrong, please let us know!

Scaffold can do lots more. Just let us know what you want. Or simply send us your templates, and we can convert it into re-usable scaffold.

We recently added several partial scaffolding, which you can choose to apply to the templates that're generated from scaffolding.

One of the most interesting one should be the fancy Suggest-As-You-Type. Users can start typing and some words or phrases matching what's already typed will show up as hints.

BTW: To see this feature in action, you need to configure which database column to spell check first.

Other partial scaffoldings, for example, multi-selection narrowBy, tag cloud, pagination, are relatively new. You may also want to check them out.

The scaffolding directory structure is also changed, so that you can add your own scaffolding without being overwritten when DBSight upgrades.

The strong feature of DBSight is searching databases, but because DBSight is only for database, it has been the shortcoming also.

But thing's changed! Now DBSight can search outside database also!

Basically you just need to implement a simple abstract java class, where you can retrieve content from any data source, like files on disk, XML feeds via HTTP, etc.

The API is simple, yet able to handle incremental indexing and a re-creating index. You just need to select the content and submit it to DBSight. The content is processed in a streaming fashion, without consuming too much memory. DBSight will handle uniqueness checking and content updating in the background.

Spell Checking has been in DBSight from the very beginning. However, it was based on a word file, instead of words taken from the indexes.

Now with DBSight 2.0.5 beta, with just one click from the dashboard, the dictionary can be generated for each index. So now you would have more reasonable spell check recommendations.

And you can specify which field needs spell checking. I think you don't want to spell check on numbers, or URLs, or dates. Just configure them via the new "Spell Check" menu.

Let us know if you have some suggestions or other interesting ideas.