Nightly MySQL exports vs change data capture: when to switch

A nightly export is how most teams start: dump the tables, load them into BigQuery, done. It works well until it doesn't.

How they compare

• Freshness: exports are as old as the last run; CDC is seconds behind.
• Load on MySQL: exports read every row of every table on each run; CDC reads the binary log, like a replica.
• Deletes: exports only remove deleted rows with full reloads; CDC applies them as they happen.
• BigQuery cost: full reloads rewrite whole tables; CDC writes only what changed.
• Effort: exports are scripts you maintain; managed CDC is a setup step.

Signs it's time to switch

• Dashboards are “wrong until tomorrow” and people ask for fresher data.
• The export window keeps growing and starts to overlap with business hours.
• Deleted or cancelled records still show up in reports.
• You need history: what a row looked like last week, not just now.

Moving over doesn't mean starting from scratch: a CDC pipeline copies the existing rows once, then streams changes. See how to replicate MySQL to BigQuery in real time.