Run scripts in IntelliShell

IntelliShell is the MongoDB shell inside Oxys DB Studio. You write JavaScript and shell commands in an editor, run the whole script, one statement or a selection, and see the documents in result tabs with the Table, Tree and JSON views. Each IntelliShell tab has its own shell process, so variables survive from one run to the next.

Open IntelliShell

1. Connect, then select a connection, database or collection in the connection tree.
2. Press Ctrl+L, click IntelliShell in the toolbar, or right-click and choose Open IntelliShell.

The tab is titled IntelliShell: <connection name>. Its current database is the one you selected. Opened on a collection, the editor holds that collection’s default query (db.getCollection("orders").find({})) and runs it, unless you turned that off (see Preferences below).

IntelliShell runs mongosh, which ships with the application. Nothing else needs to be installed.

Run a script

1. Type your script in the editor, for example:

   use acme_store
   db.orders.find({ status: "pending" }).sort({ orderedAt: -1 }).limit(20)
   

2. Choose how much to run:

3. Watch the progress (Executing..., with the elapsed time). To stop, click Stop the running script: the server operation is interrupted first; JavaScript that never yields is stopped by ending the shell process, and the next run starts a fresh shell.

Statements run one at a time. The first statement that fails stops the script and is reported with its line. Statements end at ; or at a line break, unless the next line continues them (a line starting with . for chained methods, for example).

Before anything runs, the script is checked: unbalanced brackets or an unterminated string or comment is reported as Syntax error at line N and nothing is sent.

Results

• Raw shell output shows the text output of every command, as the shell prints it. A cursor prints its first 20 documents; type it and run it to continue.
• Documents become result tabs (Find, Aggregate, Shell Output), with Table, Tree and JSON views. Results kept from the shell are read-only and hold at most 1,000 documents.
• Every run replaces the result tabs that are not pinned. Pin a tab (Pin Result: keep it after the next run) or click Pin All Results to keep them.

Query Assist

With Query Assist on (the default), a standalone db.<collection>.find(…), findOne(…) or aggregate([…]) statement is run by the application instead of the shell. Its result tab (Find Query (line N), Aggregate Query (line N)) is then paged and editable like a Collection Tab. A wand next to the line number marks the statements Query Assist will run. Other statements still go to the shell, in order.

Safety

Confirmation before destructive commands

A script that can update or delete documents, or drop a database, collection or index (updateMany, deleteOne, drop, dropIndex and the like) asks first. The Run Script dialog names the methods it found. Click OK to run, or Cancel to run nothing. Tick Don’t show this message again to stop asking, for every tab and connection; turn it back on with the Confirm destructive operations preference.

Scripts that build a command as they run

Some scripts only reveal a destructive command while running (a command name assembled from variables, for example). IntelliShell then stops the script before that command is sent, and asks:

• Run the rest resumes at the stopped line in the same shell. What already ran is not run again.
• If the stopped statement had already written something, it cannot be resumed partway: Run the rest skips the rest of that statement and continues with the statements after it.
• Any other run in the tab, or answering no, discards the stopped run; it can no longer be resumed.

Read-only connections

On a connection with the Read-Only Lock (see read-only connections), a script that calls a write method (insert, update, delete, drop, index or collection creation, user and role changes) is refused before anything runs: The connection ‘…’ is read-only: the script uses … and was not run. The shell itself also runs under a guard that refuses writes, spawning processes and writing files.

This check is a strong safety net, not a permission system. The guarantee that nothing is written is a database user without write privileges.

The legacy mongo shell cannot enforce the lock, so IntelliShell does not run it on a read-only connection.

Scheduled scripts run isolated

When an IntelliShell script runs as a scheduled task (see Scheduled scripts below), its shell is isolated: it cannot load code from files, start processes, write files outside its own temporary folder, or use eval, load, snippet, fetch or dynamic import. The script is checked when you schedule it, and constructs that could reach around these limits are refused then, with the reason. Scheduled runs need mongosh (the bundled one qualifies); the legacy shell is refused.

Scripts: load, save, history

Toolbar and editor reference

Preferences

Open Edit > Preferences (Ctrl+P), then Tools > IntelliShell. A connection can override the shell choice in its own IntelliShell tab (see the connection manager).

Older servers

Servers older than 3.6 are opened on the legacy mongo shell, which you install yourself (it is never bundled). See older servers.

Troubleshooting

Scheduled scripts

IntelliShell scripts can run as tasks: in the Tasks view, click New Task and choose IntelliShell Script, pick the script’s source and its connection, then give the task a schedule. Scheduled runs always use the isolated shell described above, so a script that works in an IntelliShell tab may be refused at schedule time if it relies on eval, load or file access.