If data is case-insensitive, or uses a collation other than ISO_BINENG, you must decrypt ciphertext columns to perform string comparisons.
When performing comparisons on strings, the distinction between equal and identical strings is important for many collations and depends on the CASE option of CREATE DATABASE. In a database that is set to CASE RESPECT and uses the ISO_BINENG collation, the defaults for SAP Sybase IQ, equality, and identity questions are resolved the same way.
Identical strings are always equal, but equal strings may not be identical. Strings are identical only if they are represented using the same byte values. When data is case-insensitive or uses a collation where multiple characters must be treated as equal, the distinction between equality and identity is significant. ISO1LATIN1 is such a collation.
For example, the strings “ABC” and “abc” in a case-insensitive database are not identical but are equal. In a case-sensitive database, they are neither identical nor equal.
The ciphertext created by the Sybase encryption functions preserves identity but not equality. In other words, the ciphertext for “ABC” and “abc” will never be equal.
To perform equality comparisons on ciphertext when your collation or CASE setting does not allow this type of comparison, your application must modify the values within that column into some canonical form, where there are no equal values that are not also identical values. For example, if your database is created with CASE IGNORE and the ISO_BINENG collation and your application applies UCASE to all input values before placing them into the column, then all equal values are also identical.