TABLE 1: Comparison of Defragmentation Methods
Command	Data Availability	Speed	Simplicity	Log Impact	Statistics	Additional Space Required in Filegroup
DBCC DB-REINDEX	Lowest: Data is offline for the duration of the command.	Fastest: Faster than individual INDEX DROP/CREATE statements. DBCC DBRE-INDEX can take advantage of more optimizations than individual DROP INDEX and CREATE INDEX statements can.	High: Can rebuild all indexes automatically and atomically; however, all your indexes must be fragmented.	High, but can be reduced by using Bulk-Logged recovery model.	Updated	Highest: All data sorted in data filegroup.
DBCC INDEX-DEFRAG	High: Data is not locked for long periods.	Slowest for heavily fragmented objects, but you can stop and start the command at will.	Low: You must specify individual indexes.	Low: Log can be backed up while the command is active.	Not updated	Lowest: Operates at the data-page level.
CREATE INDEX...DROP_EXISTING	Medium: Data unavailable only while the command is operating on clustered index.	Medium: Can avoid a sort and also just recreate defragmented indexes, but each each non-clustered index rebuild will scan the data.	Low: You must specify individual indexes and use transaction semantics to make the process atomic.	High, but can be reduced by using Bulk-Logged recovery model.	Updated	Medium: SORT_IN_TEMPDB option uses tempdb for intermediate sorting.