As part of the recent trend to mimic comp.sys.palmtops...
I'm looking for a relational database package to put on my 200LX, with a few specific needs:
1. Relational (obviously).
2. Able to work with tables that have a couple thousand records (at the most) without needing EMS or getting horribly bogged down.
3. Speedy, interactive data entry, without a lot of hoops to jump through. I don't want to have to write SQL queries every time I insert data, for example.
4. Relatively sophisticated table join capability.
I'm trying to analyze some data for possible cause-and-effect correlations. I've got two main categories of entries (let's just call them 'A', the causes, and 'B', the effects), with assorted subtypes in each. The 'A' and 'B' records will likely be stored in different tables, as they have different properties to be logged. What I will need to do is look at all the 'B' records, and find any 'A' record that falls within a certain time frame prior to any 'B' record (say 24 hours). These would be the candidate 'causes'.
I can get this far with just 1-2-3, though it isn't pretty. But then I need to go further, and look at all the subtypes of the 'A' records (causes) that were found, and determine the percent of times each subtype is then followed by each subtype of 'B' record (effect) within a 24-hour window. This would be to filter out the false positives and red herrings. If a possible cause is occurring almost constantly, but the effect in question occurs much less frequently, it's probably not a likely cause.
I do database administration and development for a living, so I could crank this out with SQL Server in about 10 minutes, but if I can do it entirely on the 200LX where I'm logging the data, that would save me the trouble of having to export/import the data, and do the analysis on a separate machine.
What are the best options? The obvious names I hear tossed around are dBase III Plus and Data Perfect, but having never used them, I have no idea what kind of table join capabilities they have. (I need to do fuzzy joins based on time stamps falling within ranges rather than exact matches, as well as outer/anti joins.) If I could dig up a copy of @Base, then I could do all the data entry via 1-2-3 (as I am now), and jump over to dBase occasionally to perform the analysis.