> ## Documentation Index
> Fetch the complete documentation index at: https://orangeslice.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# CSV Import & Enrichment

> Enrich imported CSV data by filling only empty values

# CSV Import & Empty Value Enrichment

A common workflow is importing a CSV with partial data and enriching only the missing values. This pattern saves API calls and avoids overwriting existing data.

<Steps>
  <Step title="Import CSV">
    Your CSV has partial data—some fields filled, some empty
  </Step>

  <Step title="Create View">
    Filter to show only rows where the target field is empty
  </Step>

  <Step title="Set Up Enrichment">
    Either convert the column to code or create a separate enrichment column
  </Step>

  <Step title="Run Empty Rows Only">
    Run only the filtered rows to enrich missing values
  </Step>
</Steps>

***

## Example Scenario

You import a CSV with contact data. Some emails are already filled in, others are missing:

| First Name | Last Name | Company    | Email                                 |
| ---------- | --------- | ---------- | ------------------------------------- |
| John       | Doe       | Acme Corp  | [john@acme.com](mailto:john@acme.com) |
| Jane       | Smith     | TechStart  |                                       |
| Bob        | Wilson    | BigCo      | [bob@bigco.com](mailto:bob@bigco.com) |
| Alice      | Brown     | StartupXYZ |                                       |

You want to enrich **only** Jane and Alice's emails without overwriting John and Bob's existing data.

***

## Two Approaches

There are two ways to handle this:

<Tabs>
  <Tab title="Approach A: Direct Column Enrichment">
    **Best for:** When you want to fill in the same column that has missing values.

    1. Create a view that filters for empty Email values
    2. Convert the Email column to an enrichment column (or it already is one)
    3. Navigate to the view and run **only those rows**

    <Warning>
      Be careful not to run rows that already have values—the view helps you target only the empty ones.
    </Warning>

    ```javascript theme={null}
    // Email column code - runs on rows you manually select
    const firstName = await ctx.thisRow.get("First Name");
    const lastName = await ctx.thisRow.get("Last Name");
    const company = await ctx.thisRow.get("Company");

    // Find and enrich contact info
    const contactInfo = await services.person.contact.get({
      firstName,
      lastName,
      company,
    });

    return contactInfo?.work_emails?.[0] || contactInfo?.personal_emails?.[0];
    ```
  </Tab>

  <Tab title="Approach B: Separate Enrichment Column">
    **Best for:** When you want to preserve the original column and check for existing values automatically.

    Create a new column (e.g., "Email (Enriched)") that checks for existing values before enriching.

    <Note>
      **Important:** A column cannot reference itself with `ctx.thisRow.get()`. You must reference a *different* column to check for existing values.
    </Note>

    ```javascript theme={null}
    // "Email (Enriched)" column - checks the original "Email" column first
    const existingEmail = await ctx.thisRow.get("Email");

    // Skip enrichment if email already exists
    if (existingEmail && existingEmail.trim() !== "") {
      return existingEmail;
    }

    // No existing email - proceed with enrichment
    const firstName = await ctx.thisRow.get("First Name");
    const lastName = await ctx.thisRow.get("Last Name");
    const company = await ctx.thisRow.get("Company");

    const contactInfo = await services.person.contact.get({
      firstName,
      lastName,
      company,
    });

    return contactInfo?.work_emails?.[0] || contactInfo?.personal_emails?.[0];
    ```
  </Tab>
</Tabs>

<Warning>
  **You cannot reference your own column.** `ctx.thisRow.get("Email")` inside
  the Email column will not work. If you need to check for existing values, you
  must use a separate enrichment column that reads from the original column.
</Warning>

***

## Creating a View for Empty Values

Create a view that shows only rows where the Email field is empty. This lets you target exactly which rows need enrichment.

<Tabs>
  <Tab title="Method 1: Filter Panel">
    1. In the right-hand panel, click **Filters**
    2. Click **Add Filter**
    3. Set the filter:
       * Column: **Email**
       * Operator: **is empty**
    4. Click **Save filters as view**
    5. Name your view (e.g., "Missing Emails")
  </Tab>

  <Tab title="Method 2: New View">
    1. In the left sidebar, right-click on your sheet
    2. Click **+ New View**
    3. Name your view (e.g., "Missing Emails")
    4. Navigate to the **Filters** tab
    5. Add a filter:
       * Column: **Email**
       * Operator: **is empty**
    6. Save the view
  </Tab>

  <Tab title="Method 3: Ask Chat">
    Ask the AI assistant in chat:

    > "Create a view called 'Missing Emails' that shows rows where Email is empty"

    The assistant will create the view with the correct filters automatically.
  </Tab>
</Tabs>

***

## Running the Enrichment

Once your view is set up:

1. **Click into the view** — Only rows with empty Email values appear
2. **Select the rows** — Choose which rows to enrich (or select all in the view)
3. **Click "Run"** — Execute the enrichment on selected rows only
4. **Watch progress** — As emails are enriched, rows disappear from the view (they no longer match the "is empty" filter)

This gives you real-time progress feedback and ensures you never accidentally overwrite existing data.

***

## Complete Example: Approach B

Here's a full example using the separate enrichment column approach:

```javascript theme={null}
// ============================================================================
// COLUMN: "Email (Enriched)"
// ============================================================================
// This column checks the original "Email" column first, then enriches if empty.

// Step 1: Check if email already exists in the original column
const existingEmail = await ctx.thisRow.get("Email");

if (existingEmail && existingEmail.trim() !== "") {
  // Email exists - return it without making any API calls
  return existingEmail;
}

// Step 2: No existing email - gather data for enrichment
const firstName = await ctx.thisRow.get("First Name");
const lastName = await ctx.thisRow.get("Last Name");
const company = await ctx.thisRow.get("Company");

// Step 3: Find contact information
const contactInfo = await services.person.contact.get({
  firstName,
  lastName,
  company,
});

// Step 4: Return the best email found
return contactInfo?.work_emails?.[0] || contactInfo?.personal_emails?.[0];
```

**Result:**

| First Name | Last Name | Company    | Email                                 | Email (Enriched)                                  |
| ---------- | --------- | ---------- | ------------------------------------- | ------------------------------------------------- |
| John       | Doe       | Acme Corp  | [john@acme.com](mailto:john@acme.com) | [john@acme.com](mailto:john@acme.com)             |
| Jane       | Smith     | TechStart  |                                       | [jane@techstart.com](mailto:jane@techstart.com)   |
| Bob        | Wilson    | BigCo      | [bob@bigco.com](mailto:bob@bigco.com) | [bob@bigco.com](mailto:bob@bigco.com)             |
| Alice      | Brown     | StartupXYZ |                                       | [alice@startupxyz.io](mailto:alice@startupxyz.io) |

The enrichment column returns existing values for John and Bob (no API calls), and enriches only Jane and Alice.

***

## Workflow Summary

| Step           | What Happens                                  | Why It Matters               |
| -------------- | --------------------------------------------- | ---------------------------- |
| 1. Import      | User imports CSV with some empty values       | Starting point               |
| 2. Create View | Filter `WHERE Email IS EMPTY`                 | Shows only rows needing work |
| 3. Set Up Code | Either convert column or create separate one  | Prepares the enrichment      |
| 4. Run in View | Run only the filtered/selected rows           | Targeted enrichment          |
| 5. Progress    | Enriched rows leave view as they're filled in | Visual feedback              |

***

## Best Practices

<AccordionGroup>
  <Accordion title="Cannot Reference Your Own Column">
    A column cannot use `ctx.thisRow.get()` to read its own value. If you need to check for existing data before enriching, create a **separate** enrichment column that reads from the original column.

    ```javascript theme={null}
    // ❌ WRONG: Inside "Email" column
    const existingEmail = await ctx.thisRow.get("Email"); // Won't work!

    // ✅ CORRECT: Inside "Email (Enriched)" column
    const existingEmail = await ctx.thisRow.get("Email"); // Reads from different column
    ```
  </Accordion>

  {" "}

  <Accordion title="Use Views to Target Empty Rows">
    Create views for "needs enrichment" states. Navigate to the view before
    running to ensure you only process rows that actually need enrichment.
  </Accordion>

  {" "}

  <Accordion title="Batch Wisely">
    Select manageable batches (50-100 rows) to run at a time. Monitor for errors
    before processing all data.
  </Accordion>

  <Accordion title="Prioritize Work Email">
    When enriching contact info, prioritize work emails over personal for B2B outreach:

    ```javascript theme={null}
    const contactInfo = await services.person.contact.get({ firstName, lastName, company });

    // Prioritize work email
    return contactInfo?.work_emails?.[0] || contactInfo?.personal_emails?.[0];
    ```
  </Accordion>

  <Accordion title="Handle Missing Input Data">
    Always validate that you have the required input data before making API calls:

    ```javascript theme={null}
    const firstName = await ctx.thisRow.get("First Name");
    const lastName = await ctx.thisRow.get("Last Name");

    if (!firstName || !lastName) {
      return; // Can't enrich without name
    }

    // Proceed with enrichment...
    ```
  </Accordion>
</AccordionGroup>