Code Panels

Code panels allow you to run code in a few major languages: JavaScript, Deno, Ruby, R, Julia, Python, PHP, and SQL.

Like every other panel, code panels must have a result. You can set the result using the DM_setPanel function. And you can fetch another panel's result using the DM_getPanel function.

SQL code panels are the exception. Their result is set automatically to the result of their SQL query. You cannot use DM_setPanel in a SQL code panel.

Fetching another panel's result

All code panels can grab another panel's result with the DM_getPanel function. This function takes one argument which is either the integer index of the panel (from top to bottom starting with 0), or the panel's name.

Let's say you have a file panel that reads a CSV file from disk (the Hudson River Plant dataset). And say the panel's name is Load plant data.

Once you have run the file panel, you can grab that data in a code panel using DM_getPanel('Load plant data'):

Setting a panel result

DM_setPanel sets the current panel's result. It accepts one argument. The argument must be serializable. Serialization happens automatically. Some kinds of objects that aren't serializable are circular data structures.

Building off the last example, the whole code panel should like like this:

Result format

Panel results can be of any format, as long as the value is serializable. But when setting a panel you intend to have read from a Table, Graph, or panel that supports SQL, there is a single accepted data format.

Table and Graph panels can only read from a panel that formats data as an array of objects. Each object-element must be a mapping of column name to column value. For example:

[
  { "age": 1, "name": "Eliza" },
  { "age": 3, "name": "Javier" }
]

Modifying the default program implementation

For most languages, DataStation looks in your $PATH environment variable for an implementation of the language you pick in the code panel. The SQL option is an exception, you don't need anything installed to use it.

If an implementation by the name it looks doesn't exist, the panel will fail with an error when you run the panel.

But maybe you have multiple instances of python on your machine. You can specify an absolute path you want to use for each language implementation in settings under the "Language Path Overrides" section.

In this screenshot you can see the Deno, Julia, R, and Ruby language paths are not absolute. They will be looked up in $PATH. However, JavaScript, PHP and Python are absolute paths that will not be looked up in $PATH.

Changes take effect the next time you run the code panel.

If you ever mess up these settings, you can always revert back to the default by hitting the "Reset button".

Importing 3rd party libraries

DataStation uses the language implementation you have already installed on your system. This means you can import any libraries you already have installed too.

Furthermore, this means that even when DataStation doesn't support a particular API or database, you can install a language-level package on your system so you can write your own custom code to make requests or queries to any system.

SQL

SQL code panels provide a way of running SQL queries solely against other panels, without the need for any external database or data sources configured. You don't need to set up a data source to use it. It is useful in situations where you don't have a database but the most natural way to express a join or filter is in SQL rather than code.

Under the hood this is implemented with SQLite.

SQL panels can refer to the results of other panels by using DM_getPanel($number) where $number is the zero-indexed panel you want to grab. Or, like other languages, you can also refer by name DM_getPanel('$panelName') where $panelName is the name of the panel you want to grab.

The panel you refer to must have a result that is in the array of objects format.

Building off the above plant example, here's how you could calculate most common plants with a SQL code panel.