Skip to main content
The Operations tab on a single table is where the maintenance commands live. Open a table in phpMyAdmin (left tree → database → table), then click Operations at the top.
Operations tab with Move, Rename, Order by, Table options, Copy, Maintenance

​
Repair a corrupted table

Symptoms: queries error with “table is marked as crashed”, the table won’t open, or rows are reading back as garbage.
1

Click Repair table under Maintenance

Bottom-right of the Operations page, in the Table maintenance group.
2

Wait

For small tables it’s instant. For tables with millions of rows it can take a few minutes.
3

Read the result

Successful repair shows OK under the Msg_text column. If it doesn’t, check the common issues below.
InnoDB tables don’t need repair. The engine handles its own crash recovery automatically. The Repair button is a no-op for InnoDB. It’s only useful for legacy MyISAM and Aria tables.

​
Optimize a table

Optimize rebuilds the storage layout, freeing up space left behind by deleted rows. Worth running:
  • After a big bulk delete
  • On heavily-updated tables once a quarter
  • Never on InnoDB unless you’re tight on disk space
Click Optimize table under Maintenance. It locks the table for the duration, so don’t run it on a busy production table during peak hours.

​
Rename a table

1

Find the Rename table to section

Top of the Operations page, in the Table options column.
2

Type the new name

The new name keeps the database prefix you’re already in.
3

Click Go

The rename is atomic for InnoDB. Any application or query referencing the old name will start erroring immediately.
Renaming a table the live site uses takes the site down until you update the code. Coordinate with deploys, or take a JetBackup snapshot first.

​
Change collation

Collation controls how text columns sort and compare, especially for non-ASCII characters. Change the table-level default in Table options → Collation, then click Go. To change collation on existing columns (not just new ones), tick Change all column collations in the same section. The most common change: legacy WordPress installs on utf8_general_ci, moving to utf8mb4_unicode_ci for proper emoji support.

​
Truncate (empty without dropping)

TRUNCATE deletes every row but keeps the table structure. Faster than DELETE FROM ... for big tables because it doesn’t go row-by-row.
1

Find Empty the table (TRUNCATE) under Delete data or table

Bottom-left of Operations.
2

Click it and confirm

A confirmation dialog appears. Type the table name to confirm.
3

Wait

Truncate is near-instant on most tables. Auto-increment counters reset to the starting value.
Truncate is irreversible. There’s no rollback even inside a transaction (in MySQL, TRUNCATE is implicitly committed). Take an export of the table first.

​
Drop a table

Drop removes the table and all its data permanently.
1

Find Delete the table (DROP)

Below Truncate, in the same red-tinted box.
2

Click and confirm

Type the table name to confirm. The table is gone.
No undo. If it’s a production table, take a JetBackup snapshot or a phpMyAdmin export first. The 30-second cost up front is much cheaper than the recovery you’d otherwise be doing.

​
Copy a table

Useful for staging changes. Copy orders to orders_backup_20260427, run your migration on the live orders, and you have a fast rollback target.
1

Find Copy table to under Move/Copy

Right side of the Operations page.
2

Pick database and new name

The copy can stay in the same database or land in another. Same database + a new name is the common case.
3

Pick what to copy

  • Structure only: empty copy with the same columns and indexes
  • Data only: copy rows into an existing table (rare)
  • Structure and data: full copy
4

Click Go

The copy completes. The new table appears in the left tree.

​
Common issues

The table is too large to repair on the available disk space (repair needs to write a temp copy alongside the original). Either increase disk through your plan, or open a ticket and we’ll repair it from the server side where we have more headroom.
File-system permissions on the database directory got out of sync, usually after a botched restore. Fixing this needs root access, so open a support ticket.
That’s expected. InnoDB doesn’t have an in-place optimize, so the engine does a full rebuild plus statistics refresh. Functionally equivalent to optimize, just longer.

​
Need a hand?