Does C&A gathering store the entire message body?
When configuring a Content and Attachment gathering and specifying a body keyword, MessageStats will store a record for each message that matches that keyword. Each of these records contains a number of statistics including the message size, subject line, recipients, etc. Additionally MessageStats stores a record of the search keywords that were found in each message. This allows users to generate an Audit Items by Body Keyword report. It is important to note that MessageStats does not store the body of the message itself, only the search keywords that were found in the message.
If the gathering is configured to collect All content, then MessageStats will store a record for EVERY message in the mailbox. Again, these records contain general information about each message; the size, the subject, originator, recipients, etc. As previously noted, MesageStats does not store the body of the message, and since users are not looking for any specific keywords in the message body MessageStats does not store any information about the words that were found in the message. In fact in this case MessageStats does not look at the message body at all during the gathering.
Basically MessageStats does store the body of the message, but only if it is queried for specific keywords. If no query is done for keywords MessageStats does not store the body of the message.
This is by design. It is unlikely that there would be a need to ever store the entire message body for every message in the mailbox. This could potentially use up a lot of space in the MessageStats database.