Bookshelf Home | Contents | Index | Search | PDF |
Siebel Marketing Guide > Defining Attributes and Buckets > Defining Attribute Families >
Trapping Unclean Data
Even the best data warehouse can have values that are outside the expected range. This is referred to as unclean data. In the MARITAL_STATUS example, possible values are 1 through 4. If the data is corrupt, values such as -2, -4.5, or 297 may exist.
To trap unclean data, you can create entries to trap the bad values, as shown in Table 32.
Table 32. Creating Entries to Trap Data Label Code Operator Value Single S = 1 Married M = 2 Divorced D = 3 Widowed W = 4 Negative Trap NT <= 0 Positive Trap PT >= 0When you review your data, check the traps for data. If data is being captured, you should check the source data to locate and correct the unclean data.
Remember, data changes over time, and you may not have data in the traps the first time you extract information.
Bookshelf Home | Contents | Index | Search | PDF |
Siebel Marketing Guide, Version 7.5, Rev. A Published: 18 April 2003 |