grant@NRL-CSS.ARPA (William (Liam) Grant) (08/12/87)
Hello all: I already know that my use of Datatrieve will have less than optimum efficiency because of the length of my records. I am looking for suggestions within the constraints set by my employers. These include using Datatrieve, with indexed fields, and creating the fields that they have determined are necessary, regardless of my opinion. In addition to setting fields, they have specified minimum field widths in most cases. Lastly, these structures must be accessible from procedures written in Datatrieve (so that a clerk/typist doesn't have to learn Datatrieve). I've solved most of the specs, but I have one last question concerning efficiency. It concerns record size. They are already 125 bytes each, spread over several smaller fields. I must now add a new field of at least 100 bytes. According to what I remember about a file structures course, using an indexed structure is more efficient when an integer number of records fits in a disk block. Using RA81's, this would mean records of either 128 or 256 bytes would be more efficient than 225 bytes. Is this true ? Have I forgotten something ? Have I missed something in the manuals ? (With so many for ONE layered product, it's easy). Reply to me and I'll summarize. grant@nrl-css.arpa grant@nrl-com.arpa (202) 767-2392 William (Liam) Grant Naval Research Laboratory Code 5522 Washington, D.C. 20375-5000