Development & IT Consultation

Service frequently asked questions

• Will the same SQL + DDL run against Postgres/MySQL and BigQuery ?

  Often yes, but with disastrous cost in BigQuery, or disastrous performance in MySQL or Postgres.

• How can a bad SQL affect the MySQL, Postgres ?

  In MySQL and Postgres, and bad SQL can bring the server to a halt, where the server stops responding to all requests until rebooted & cleaned up.

• How can a bad SQL affect the BigQuery?

  In BigQuery, a bad SQL can cost a LOT. I've been presented with $1M/month BigQuery bills, and changed the approach with SQL and DDL to do the same for pennies.

• Can all SQL be improved ?

  I've never seen a SQL statement I couldn't improve somehow. There is always a way.

• What is something everyone should know in SQL evening beginners

  AVOID cartesian joins. ESPECIALLY on big data. There is always a better way.

• What is something people don't know is bad in SQL but should

  Don't use the NOT IN clause. The work the database does to resolve this is excessive.
  There is always a better way.

• What cost performance improvement can I expect in BigQuery

  Every single query against BigQuery should be designed with the goal to be under the minimum byte scan charge of 10MB, which is fractions of a cent. Even when querying a 20GB table, your individual queries should be fractions of a cent.

• For an operational database, is BigQuery best or MySQL or Postgres

  NEVER using BigQuery for and transactional data activities. the updates are slow, limited in number per day, and expensive.

  Postgres is best, MySQL good second choice.

• Best database for mining log files

  Bigquery was built for this. it's amazing. Just watch your cost.