# CSV file format

{% hint style="info" %}
CSV File Uploads **only work** for custom integrations that can map their resources to Tables and Columns. Lineage is not support for CSV File Upload at this time.
{% endhint %}

Generate a CSV of your data that contains a list of all of your columns in the database in the following format.

| col\_name | col\_description | col\_type | sort\_order | database | schema | name  | description            | is\_view | last\_updated\_time |
| --------- | ---------------- | --------- | ----------- | -------- | ------ | ----- | ---------------------- | -------- | ------------------- |
| id        | ID of the user   | int       | 1           | secoda   | public | users | the table of all users | false    | None                |
| name      | name of the user | char      | 2           | secoda   | public | users | the table of all users | false    | None                |

* **col\_name** -> The name of the column (ex. `event_name`)
  * :exclamation:**Required**
  * Column names can be made up of numbers, letters, and any special characters **except** periods (ex. `event.name` :thumbsdown:).
  * :information\_source: Column names will be made lower case when brought into Secoda
* **col\_description** -> The description of the column
  * Column descriptions can be made up of numbers, letters, and any special characters.
  * :information\_source: There is no additional validation or transformation on this field. The way you input it into the CSV is the way it will show in the UI.
* **col\_type** -> The type of data in the column (ex. `character varying(25)`)
  * Column types can be made up of numbers, letters, and any special characters.
  * :information\_source: There is no additional validation or transformation on this field. The way you input it into the CSV is the way it will show in the UI.
* **col\_sort\_order** -> The default order that the columns should be sorted in
  * ❕**Sort order must be an integer.**
  * :information\_source: If sort order is not important, you can set the field to `0` and it will default to alphabetical order.
* **name** -> The name of the *table* the column belongs to
  * :exclamation:**Required**
  * Table names can be made up of numbers, letters, and any special characters **except** periods.
* **description** -> The description for the *table* the column belongs to.
  * Table descriptions can be made up of numbers, letters, and any special characters.
  * :information\_source: There is no additional validation or transformation on this field. The way you input it into the CSV is the way it will show in the UI.
* **database** -> The name of the database the column and table belong to.
  * :exclamation:**Required**
  * Column names can be made up of numbers, letters, and any special characters **except** periods.
* **schema** -> The name of the schema the column and table belong to.
  * :exclamation:**Required**
  * Column names can be made up of numbers, letters, and any special characters **except** periods.
* **is\_view** -> Indicates whether the *table* is a view (generated from a query).
  * Is View is a boolean value expecting either `true` or `false`.
* **last\_updated\_time** -> The timestamp of when the *table* was last updated.
  * Last updated time is a timestamp that expects an integer value of the Unix Epoch time.

{% hint style="info" %}
For rows which contain columns belonging to the **same** table, the `is_view`, `last_updated_time`, `name`, and `description` fields should all be identical.

Ex. If there are 5 columns all belonging to the table `TestTable`, each row for the columns in the CSV would have a name of `TestTable` and an identical description.
{% endhint %}

For example with Snowflake, this CSV can be generated with a SQL statement like this.

```sql
SELECT DISTINCT
    lower(c.column_name)                                    AS col_name,
    c.comment                                               AS col_description,
    lower(c.data_type)                                      AS col_type,
    lower(c.ordinal_position)                               AS col_sort_order,
    lower(c.table_catalog)                                  AS database,
    lower(c.table_schema)                                   AS schema,
    lower(c.table_name)                                     AS name,
    t.comment                                               AS description,
    decode(lower(t.table_type), 'view', 'true', 'false')    AS is_view,
    DATE_PART(EPOCH, t.last_altered)                        AS last_updated_time
FROM {database.title}.INFORMATION_SCHEMA.COLUMNS AS c
LEFT JOIN {database.title}.INFORMATION_SCHEMA.TABLES t ON c.TABLE_NAME = t.TABLE_NAME
 AND c.TABLE_SCHEMA = t.TABLE_SCHEMA
WHERE c.TABLE_SCHEMA not in ({','.join(SnowflakeConnectionConstants.DEFAULT_IGNORED_SCHEMAS)}) \
        AND c.TABLE_SCHEMA not like 'STAGE_%' \
        AND c.TABLE_SCHEMA not like 'HIST_%' \
        AND c.TABLE_SCHEMA not like 'SNAP_%' \
        AND lower(c.TABLE_SCHEMA) not like 'dbt_cloud_pr%' \
        AND lower(c.TABLE_SCHEMA) not like 'ci_impact_resilience_%' \
        AND lower(c.COLUMN_NAME) not like 'dw_%' \
```

Feel free to reach out to us for more guidance on how to generate CSVs.

At this point, you cannot add custom properties or tags using the CSV upload. However, you can add custom properties and tags after the initial extraction has been done using the [Import/Export ](/resource-and-metadata-management/import-and-export-data.md)feature.


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://docs.secoda.co/integrations/custom-integrations-and-marketplace/custom-integration/create-your-csv.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.