Analyzing the problem

Before you start, you must understand the reporting problem that you are trying to solve and what data is available to solve it.

If you cannot address the following questions, talk to your users about their reporting requirements:


  1. Do you and your users agree on the reporting requirements?

    Issues to resolve can include multilingualism, performance, security, and how to organize and combine query items and filters.

  2. Does the data source contain the data and metadata that you need?

    Without metadata such as primary keys, indexes, and foreign keys, your reports may take too long to run, or may produce incorrect results. If the data source does not contain the data and metadata that you need, will it be changed, or will you work around it?

  3. Does the same data exist in more than one source?

    If so, choose the data source that most closely fits your reporting requirements. If a data warehouse is available, it is typically a better choice than an operational database. A data warehouse based on a star schema is ideal. If this does not exist, and you expect that your reporting application will be heavily used, consider arranging for one to be created.

  4. Which data source tables are the fact tables, which are the dimensions, and which are both fact table and dimension?
  5. What are the keys and attributes of each dimension?
  6. Which relationships are required?
  7. Are there multiple relationship paths between tables?

    If so, what does each path represent? You must define the preferred path for each.


Then you should review the names of data sources, tables, and columns in your data source to ensure that you are not using names reserved by IBM® Cognos®. If you must use a reserved word, enclose the word in quotes in the SQL specification. For example, select Orderdate, "Timezone". For more information, see Reserved words.