README.md

Snowflake Import Plugin (BETA)

Looking to contribute a transformation? See Contributing a transformation.

Import data from a Snowflake table in the form of PostHog events.

⚠️ Important Notice

This plugin is still in Beta! Use it at your own risk. Feel free to check out its code and submit feedback.

Installation Instructions

1. Select a Snowflake table to use for this plugin

2. Create a user with sufficient priviledges to read data from your table

We need to create a new table to store events and execute INSERT queries. You can and should block us from doing anything else on any other tables. Giving us table creation permissions should be enough to ensure this:

CREATE USER posthog WITH PASSWORD '123456yZ';
GRANT CREATE ON DATABASE your_database TO posthog;

3. Add the connection details at the plugin configuration step in PostHog

4. Determine what transformation to apply to your data

This plugin receives the data from your table and transforms it to create a PostHog-compatible event. To do this, you must select a transformation to apply to your data. If none of the transformations below suit your use case, feel free to contribute one via a PR to this repo.

Contributing a transformation

If none of the transformations listed below suits your use case, you're more than welcome to contribute your own transformation!

To do so, just add your transformation to the transformations object in the index.ts file and list it in the plugin.json choices list for the field transformationName.

A transformation entry looks like this:

'<transformation name here>': {
    author: '<your github username here>',
    transform: async (row, meta) => {
        /* 

        Fill in your transformation here and
        make sure to return an event according to 
        the TransformedPluginEvent interface:

        interface TransformedPluginEvent {
            event: string,
            properties?: PluginEvent['properties']
        }

        */
    }
}

Your GitHub username is important so that we only allow changes to transformations by the authors themselves.

Once you've submitted your PR, feel free to tag @yakkomajuri for review!

Available Transformations

default

The default transformation looks for the following columns in your table: event, timestamp, distinct_id, and properties, and maps them to the equivalent PostHog event fields of the same name.

Code

async function transform (row, _) {
    const { timestamp, distinct_id, event, properties } = row
    const eventToIngest = { 
        event, 
        properties: {
            timestamp, 
            distinct_id, 
            ...JSON.parse(properties), 
            source: 'snowflake_import',
        }
    }
    return eventToIngest
}

JSON Map

This transformation asks the user for a JSON file containing a map between their columns and fields of a PostHog event. For example:

{
    "event_name": "event",
    "some_row": "timestamp",
    "some_other_row": "distinct_id"
}

Code (Simplified*)

*Simplified means error handling and type definitions were removed for the sake of brevity. See the full code in the index.ts file

async function transform (row, { attachments }) {            
    let rowToEventMap = JSON.parse(attachments.rowToEventMap.contents.toString())

    const eventToIngest = {
        event: '',
        properties: {}
    }

    for (const [colName, colValue] of Object.entries(row)) {
        if (!rowToEventMap[colName]) {
            continue
        }
        if (rowToEventMap[colName] === 'event') {
            eventToIngest.event = colValue
        } else {
            eventToIngest.properties[rowToEventMap[colName]] = colValue
        }
    }

    return eventToIngest
}