A mechanism is provided for identifying suggestions as to which log files associated with data in a data warehouse to search for particular data. Features from the contents of a plurality of log files from a set of log files are selected. The plurality of log files are grouped based on the selected features. Using extracted log event sequences, a correlation between the log event sequences and a plurality of data transaction tables from a set of data transaction tables in the data warehouse is calculated. Suggestions as to which log files in the set of log files should be searched is then identified for particular data based on a business relevance value and a utilized data ratio.