The most cost-effective way to parse JSON into columns in Clay is to create child columns from your JSON data. Click on the cell containing your JSON data to open the Cell details panel, then hover over specific endpoints and click "Add as column" to map individual fields into separate columns.
For more complex JSON parsing, you can use our "write to table" feature to parse arrays of objects into individual rows. This allows you to specify which objects to write to a new table and map them to columns.
You can also use the Formulas enrichment feature for splitting combined data into multiple columns by selecting "Add Column" and choosing "Formulas," then specifying your extraction operation.
Could you copy and paste your JSON into the chat so I can provide more specific guidance on the best parsing approach for your data structure?
You're right! For JSON parsing, the cheapest GPT model (GPT-3.5) works perfectly and costs much less than GPT-4 variants.
Use ChatGPT enrichment with GPT-3.5 model and enable "Return as JSON Object" mode. In your prompt, clearly specify which JSON fields you want extracted and mention "json" in the prompt.
Alternatively, use Claygent with GPT-3.5 (default model) - it costs only 3 credits per row compared to GPT-4 Turbo which is 20x more expensive.
Both approaches let you define column outputs to automatically structure the parsed data into separate columns.
.png){kind=logo}
