Tables and SQL

Tables, under Library in the sidebar, lists every table your sources sync, by company and tool. Each row shows the table's row count, when it last synced, its status and the datasets that read it.

A table's plain name sits over its raw name, the one SQL and agents use. Beside each company is its schema, the part of a raw name before the dot, which you type in SQL.

Open a table to read what it holds, and for the rest:

• Rows by connection: how many rows each connection sent.
• Used by: the datasets that read the table.
• Columns: each column's name and type.

Then choose Open in SQL to query it. Search finds a table by its name or by a column in it.

If your workspace loads into your own warehouse, Gatlas cannot list its tables here. Browse them in your warehouse instead.

A table reads Out of date when its newest rows are 7 days or more older than its tool's last sync. The tool ran, and this table got nothing new. Agents read the same rule.

Choose Out of date in the status filter to see only those tables. Often the reason is in the tool itself, such as a ledger nobody has posted to. If it is not, open the connection on Connections, under Sources in the sidebar.

A connection has its own rule: No sync in 2 days means the tool has not synced at all for 48 hours. See Connecting sources.

SQL, under Library in the sidebar, is where you check a number at its source. It reads the synced tables of the companies you can see.

1. Find a table on the Tables tab, by company and tool and named as on the Tables page, or open one from that page with Open in SQL.
2. Write a query and choose Run, or press Cmd+Enter.
3. Choose Export CSV to take the result away.

History holds your recent queries, and Templates holds ready-made ones to start from.

When a query is worth keeping, Save as... turns it into a dataset or a draft. Your datasets are listed beside the editor, so you can open one to change its query. What a dataset is, and how to describe it for agents, is in Datasets.