Scripting the shell

The MongoDB shell is not just an interactive shell, it can also be scripted using JS files. In addition to specifying a Javascript file (*.js) you can also use --eval with a snippet of JS.

Using the shell this way allows for tasks to be performed without the need for any additional drivers or language support; it can be used in cron, or automated administrative tasks. Please be aware there are data format issues in javascript so you should be careful how much you do in Javascript.

Common uses for the scripted shell includes:

• backups
• scheduled Map-Reduce commands
• offline reports
• administration

Running a Script

./mongo server:27017/dbname --quiet my_commands.js

The syntax stems from the interactive shell. This command will execute the my_commands.js as if it had been entered into the shell directly, with some exceptions.

• ./mongo: command to start the interactive shell, may vary on your shell of choice
• server:27017/dbname: basic connection information
• --quiet: this is a flag for the mongo command. This switch removes some header information that is not typically necessary when building unattended scripts.
• my_commands.js: a file containing a series of shell commands to execute

--quiet

This option will remove the header printed at the top when you run the shell normally:

$ mongo
MongoDB shell version: 2.1.0
connecting to: test
> ^C
bye
$ mongo --quiet
> ^C

--eval

In addition to using a full Javascript file you can also pass in a Javascript fragment:

bash-3.2$ ./mongo test --eval "printjson(db.getCollectionNames())"
MongoDB shell version: 1.8.0
connecting to: test
[
        "system.indexes",
        "t1",
        "t2",
        "test.fam",
        "test1",
        "test11",
        "testBinary",
        "testarray"
]

Differences between scripted and interactive

Printing

When using the shell interactively, the shell will print returned values and format where possible. This is done as a general convenience from within the shell. However, when building a script, the printing needs to be defined explicitly.

There are two functions commonly used for this:

1. print(): works as normal javascript
2. printjson(): prints a nicely formatted JSON representation of the given object

Example: print JSON for the first 10 objects from a find

db.foo.find({x:1}).forEach(printjson)

use dbname

This command does not work in scripted mode. Instead you will need to explicitly define the database in the connection (/dbname in the example above).

Alternately, you can also create a connection within the script:

> help connect // for more help
> var x = new Mongo('host[:port]');
> var mydb = x.getDB('mydb');
> // or
> var mydb = connect('host[:port]/mydb');

it

The iterator command it does not work outside of the interactive scripting environment.

getLastError

When running an update/insert command from the shell, the shell automatically awaits a reply (i.e. runs a get last error).

The same is not true when running from a script file. To wait for the status of an operation (such as a write), run the getLastError function after update/insert.

db.getLastErrorObj()
// or
db.getLastError()