Initial Explorations

Keeping the total bandwidth for initial explorations low

To get familiar with a new database, roaming around and looking at the data is necessary. Unfortunately, large queries can slow down performance. Worse, large exports of data from a new user can flag your account as fraudulent by a SecOps AI. To avoid that, some of best practices are...

Look at Small Data

Most of the time, you don't need a full table to get an idea of what the table is holding. Make use of the LIMIT operation.

SELECT * FROM table1 LIMIT 10;

Once you're relatively sure of the columns you'll need to use in the future, the next step is determining size.

SELECT COUNT(col1) FROM table1;
SELECT COUNT(*) FROM table1;

This confirms whether the table will be the right size to pull down, or a further trim down with a WHERE statement would be more helpful. Especially before looking at more interesting ways to directly pull data, know the size of the data on which SQL is being performed.

Understand the Scope

With good business requirements, data pulls can be small and actionable. Pulling an extra month of data in each direction can give a report more context, but if your source has multiple years of data then pulling all dates for a monthly update is overkill.

SELECT min(timestamp), max(timestamp) FROM table1;

Then, trim the data to the dates needed.

SELECT col1, col2, col3 FROM table1 t
WHERE '2020-01-01' < t.timestamp < '2020-05-31'