You can create an index on an encrypted column if the column’s encryption key does not specify the use of an initialization vector or random padding. Using an initialization vector or random padding results in identical data encrypting to different patterns of cipher text, which prevents an index from enforcing uniqueness and from performing equality matching of data in cipher text form.
Indexes on encrypted data are useful for equality and nonequality matching of data but not for data ordering, range searches, or finding minimum and maximum values. If Adaptive Server is performing an order-dependent search on an encrypted column, it cannot execute an indexed lookup on encrypted data. Instead, the encrypted column in each row must be decrypted and then searched. This slows data processing.
