> ## Documentation Index
> Fetch the complete documentation index at: https://www.landbase.com/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# How to enrich an uploaded dataset end to end

> Chain multiple Landbase dataset workflows together to take a raw CSV from upload to a fully enriched, downloadable output file.

# How to enrich an uploaded dataset end to end

This guide shows you how to take a raw CSV of company or contact records, upload it to Landbase, run a sequence of dataset workflows to match and enrich the records, and download the finished output.

This is the right approach when you have your own data (a CRM export, a spreadsheet, a scrape) and want Landbase to match records to its database and add enrichment fields.

***

## Prerequisites

* landbase-cli installed and authenticated
* A CSV or Excel file with at least one identifier column (e.g. company name, domain, or LinkedIn URL)

***

## Step 1: Upload your file

```bash theme={null}
DATASET_ID=$(landbase-cli upload ./crm-export.csv --name="CRM Export Jun 2026" | jq -r .id)
echo "Dataset ID: $DATASET_ID"
```

The `upload` command returns a JSON object with an `id` field. Capturing it in `$DATASET_ID` avoids having to copy-paste it for every subsequent command.

***

## Step 2: Run the onboard workflow

The `onboard` workflow normalizes your data, maps columns to Landbase's schema, and prepares it for downstream workflows.

```bash theme={null}
landbase-cli workflow onboard $DATASET_ID --wait
```

`--wait` blocks until the workflow completes. For large datasets this can take a few minutes.

***

## Step 3: Run the match workflow

The `match` workflow compares each row against the Landbase database and adds a confidence score and a Landbase record ID to matched rows.

```bash theme={null}
landbase-cli workflow match $DATASET_ID --wait
```

***

## Step 4: Run the enrich workflow

The `enrich` workflow pulls additional fields (industry, employee count, HQ location, LinkedIn, etc.) for matched records.

```bash theme={null}
landbase-cli workflow enrich $DATASET_ID \
  --company-fields=industry,size_range,hq_city,hq_country \
  --wait
```

To see what fields are available for enrichment, run:

```bash theme={null}
landbase-cli enrich --help
```

***

## Step 5: Publish and find the output dataset

The `publish` workflow generates a downloadable file from the dataset.

```bash theme={null}
landbase-cli workflow publish $DATASET_ID --format=csv --wait
```

After publishing, the output is a **child dataset** — a new dataset linked to your original. Find it with:

```bash theme={null}
CHILD_ID=$(landbase-cli datasets lineage $DATASET_ID --direction=children \
  | jq -r '[.[] | select(.workflow_type == "publish")] | .[0].id')
echo "Output dataset: $CHILD_ID"
```

***

## Step 6: Download the output

```bash theme={null}
landbase-cli datasets download $CHILD_ID ./enriched-crm.csv
```

Your enriched file is now saved locally.

***

## Full pipeline as a script

```bash theme={null}
#!/bin/bash
set -e

FILE=${1:-crm-export.csv}
NAME=${2:-"CRM Export"}

echo "Uploading $FILE..."
DATASET_ID=$(landbase-cli upload "$FILE" --name="$NAME" | jq -r .id)
echo "Dataset: $DATASET_ID"

echo "Onboarding..."
landbase-cli workflow onboard $DATASET_ID --wait

echo "Matching..."
landbase-cli workflow match $DATASET_ID --wait

echo "Enriching..."
landbase-cli workflow enrich $DATASET_ID \
  --company-fields=industry,size_range,hq_city,hq_country \
  --wait

echo "Publishing..."
landbase-cli workflow publish $DATASET_ID --format=csv --wait

echo "Finding output..."
CHILD_ID=$(landbase-cli datasets lineage $DATASET_ID --direction=children \
  | jq -r '[.[] | select(.workflow_type == "publish")] | .[0].id')

echo "Downloading..."
landbase-cli datasets download $CHILD_ID ./enriched-output.csv

echo "Done. Output: enriched-output.csv"
```

Save this as `enrich-pipeline.sh`, make it executable with `chmod +x enrich-pipeline.sh`, and run it with:

```bash theme={null}
./enrich-pipeline.sh my-contacts.csv "My Contacts Jun 2026"
```

***

## Troubleshooting

**Workflow fails at onboard:** Check that your CSV has recognizable column headers (name, domain, email, first\_name, last\_name, etc.). The onboard step maps your columns — ambiguous headers can cause it to fail.

**No child dataset found after publish:** Run `landbase-cli datasets lineage $DATASET_ID --direction=children` without the `jq` filter to see all child datasets and their workflow types.

**Timeout:** Add `--timeout=600` to any `--wait` command to allow up to 10 minutes. Large datasets take longer.