sql server 2005 - unique constraint on large text field -
i have table column xml feed entry (nvarchar(max)) , executed hash function (md5) on column hash key aid in searching , maintain uniqueness. typically on 900 bytes long.now have realized hash key not guarantee uniqueness since did count of duplicate rows on column , amazed , when tried create unique index on hash key column itself, sql server gave me error of cannot create unique index on non deterministic column..something that.
now have searched , found in cases should use type (ntext) , create unique index on that; others mention using checksum , index on that, whilst others mention triggers, checksum, etc. best method implement on field uniqueness keeping in mind table has 22 million records , cannot afford lose data. in advance
microsoft recommends not using ntext types because deprecated in future. documentation here. go using checksums , putting unique index on that. this post discusses this.
Comments
Post a Comment