Updating MongoDB using Perl

Previously we started to build a command line phonebook. We could add new people with their phone number and list all of them or a specific person.

Let's now see how can we update an entry.

We need some way to identify the document in the database we would like to update.

First, given the name of he person, let's change the phone number:

At the top of our script we added a new attribute:

option update => (is => 'ro');

We expect the user to run the script:

perl phonebook.pl --update --name Foo --phone 789

meaning we would like to update the phone number of Foo to be 789.

The code that implements the actual change looks like this:

    } elsif ($self->update) {
        die "--name is needed" if not $self->name;
        die "--phone is needed" if not $self->phone;

        $people_coll->update(
            { name => $self->name },
            { '$set' => { phone => $self->phone }},
        );
    }

After 2 lines of very basic input validation we call the update method on the collection. It receives two hash references as parameters. The first one selects the documents that need to be updated. { name => $self->name } means to select the document with the given name.

The second has tells MongoDB what changes to make. In our case we use the $set operator that will set the phone field to the new value. (The leading $ sign is part of the dialect of MongoDB.)

That's it.

Duplicate people

By default MongoDB does not enforce uniqueness on the documents, so we can add several people called Borg:

$ perl phonebook.pl --add --name Borg --phone 123
$ perl phonebook.pl --add --name Borg --phone 456
$ perl phonebook.pl --add --name Borg --phone 789
$ perl phonebook.pl --list --name Borg
Borg  123
Borg  456
Borg  789

In general this is good, as we might have several friends with the exact same name. There is a problem with this though. First of all, when we list the information about Borg it will not be clear which one is which. We'll deal with this later.

First lets' see what happens when we try to change the phone of Borg:

$ perl phonebook.pl --update --name Borg --phone 1111
$ perl phonebook.pl --list --name Borg
Borg  1111
Borg  456
Borg  789

By default, the update method of MongoDB will only change one document. The first one it finds. In many cases this is the correct behavior, but if we would like to change the phone numbers of all the Borgs we need to tell that to MongoDB.

The update method can get a third reference to a hash with options. Setting the multiple option to be true will tell MongoDB to update all the documents that match the criteria.

        $people_coll->update(
            { name => $self->name },
            { '$set' => { phone => $self->phone }},
            { multiple => 1 },
        );

See also the MongoDB reference about update and the Perl documentation of MongoDB::Collection

$ perl phonebook.pl --update --name Borg --phone 2223
$ perl phonebook.pl --list --name Borg
Borg  2223
Borg  2223
Borg  2223

That's wonderful, except that now we have really no way to differentiate among the Borgs.

Let's start from scratch.

Delete a document

MongoDB provide the remove method. It accepts a hash reference describing one or more documents (just as in the case of update) and removes all matches.

    } elsif ($self->delete) {
        die "--name is needed" if not $self->name;
        $people_coll->remove({ name => $self->name });

We also need to add the attribute at the top:

option delete => (is => 'ro');

$ perl phonebook.pl --delete --name Borg
$ perl phonebook.pl --list --name Borg

Exercise

Now that we got rid of the old Borgs, we can change the script to let the user supply a nickname as well. In the listing we should also mention the nickname and then we can also change the part of the update to rely on both the name and the nickname.

The latest version of our script can be found in GitHub. Please fork it, and send a pull request with your solution.