Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Path: utzoo!utgpu!water!watmath!clyde!cbosgd!cwruecmp!hal!ncoast!allbery From: allbery@ncoast.UUCP Newsgroups: comp.databases Subject: Re: Database problem of the week Message-ID: <4189@ncoast.UUCP> Date: Sun, 16-Aug-87 15:46:28 EDT Article-I.D.: ncoast.4189 Posted: Sun Aug 16 15:46:28 1987 Date-Received: Tue, 18-Aug-87 00:50:17 EDT References: <923@woton.UUCP> Reply-To: allbery@ncoast.UUCP (Brandon Allbery) Followup-To: comp.databases Organization: Cleveland Public Access UN*X, Cleveland, Oh Lines: 52 As quoted from <923@woton.UUCP> by riddle@woton.UUCP (Prentiss Riddle ): +--------------- | I've inherited responsibility for a medium-large and growing database (say | ~400 Kbytes, not including index files etc.) of experimental data which needs | to be kept around and subjected to statistical analysis. | | Currently the data is stored in a single large Informix file in which each | record consists of a key and 52 floating point lab values. The key is a | composite of a study group number, a subject number and a time. | | +-----------+-------------+------+-------+-------+- -+--------+ | | group no. | subject no. | time | val 1 | val 2 | ... | val 52 | | +-----------+-------------+------+-------+-------+- -+--------+ +--------------- I'd use two tables (files): test test-id primary group-no subject-no time result test-id result-no result-id primary composite (test-id result-no) result-val +--------------- | PLAN C: This may not really be a problem for a DBMS. The kinds of queries we | tend to do are mostly simple, and we'd be perfectly happy to separate the | study groups into distinct files where they wouldn't interact at all. What I | sometimes wish I had was a simple program for manipulating a sparse 3-D | matrix with subject on one axis, time on another and variable on a third. It | would need a user-friendly screen-oriented interface for entering and | correcting data, and the ability to dump one- and two-dimensional slices of | the matrix into ascii files to be fed to statistics programs. That sounds a | lot like a 3-D spreadsheet to me, but I don't know where to find such a beast | for little or no money. Also I'm not very familiar with the internals of | spreadsheets. What sorts of data structures do they use -- hashing, perhaps? +--------------- Well, if you used PCs, you could use Qubecalc: a shareware 3-D Lotus-lookalike written in Turbo Pascal. -- Brandon S. Allbery, moderator of comp.sources.misc and comp.binaries.ibm.pc {{harvard,mit-eddie}!necntc,well!hoptoad,sun!mandrill!hal}!ncoast!allbery ARPA: necntc!ncoast!allbery@harvard.harvard.edu Fido: 157/502 MCI: BALLBERY <> ** Site "cwruecmp" is changing its name to "mandrill". Please re-address ** *** all mail to ncoast to pass through "mandrill" instead of "cwruecmp". ***