Basic SQLite question

[cag8f]

I'm an amateur programmer who has never worked with databases. I *think* that using databases (namely SQLite) can help with some of my Autoit projects, but I'm having trouble determining exactly how. I'm hoping the database literati here can help explain a particular aspect about databases that is giving me trouble.

My simplified situation is this. A user enters his height and weight into a program and, based on that height/weight combination, the program returns a group, either "A" or "B." To accomplish this using an Excel spreadsheet, I would enter every single height/weight combination, with its corresponding group. The program would read in the user data, then, using loops and conditionals, return the group.

My database question is: Can using SQLite instead of Excel be beneficial in terms of coding efficiency, processing speed, memory usage, or any other aspect? If so, how? Mainly I'm interested in saving speed, but I'm also interested to know what other aspects would be affected.

Sorry for the long, broad, vague question. I'm not looking for someone to tell me how to do something. I'm looking for someone to assure me that SQLite (rather than a simple spreadsheet) is right for me, and to hopefully steer me towards some documentation that may be beneficial to me. I posted this to the SQLite mailing list last week but apparently am still waiting for moderator approval.

Thanks in advance.

[jchd]

You question made it on the mailing list and some answers were posted there.

The summary (and my opinion as well) is that if you only have a list of 10 ranges of 2 measurements simply using an array of constants will do the job of choosing that user entries make him/her belong to group A or B.

So obviously your question is something else: do you need to store a large enough number of people data? Then SQLite might be a solution.

[cag8f]

Thanks for the reply. The last email I received from the SQLite mailing list told me that my question was pending moderator approval. How do I view replies? edit: nevermind, I see the offline page.

>> do you need to store a large enough number of people data?

Yes. I presented the small, simplified scenario just to illustrate my issue. My actual scenario involves upwards of 10 criteria (e.g. height, weight, age, etc), with upwards of 200 user-entered possibilities each. I'd like to know in what ways SQLite would be advantageous in this situation. I'm having trouble ascertaining how using SQLite would be faster.

Edited April 19, 2013 by cag8f

[jchd]

If you have that many criteria then SQLite is probably the best way to go. It would also add the benefit of allowing you to store much more application data in the same container, and permit making statistics on use, changing profiles at will, etc.

Furthermore an SQLite database is portable accross any OS and any hardware platform seemlessly since every implementation uses the exact same file format.

[cag8f]

Thanks.

>> SQLite is probably the best way to go.

Right, I figured that. But I'm trying to ascertain why exactly. Would using SQLite result in faster processing times? less memory usage? Would it just be easier to implement? something else? a combination?

[jchd]

A combination. Instead of running loops thru hard-coded values in several arrays, you can get the result with a single query. Certainly a much flexible design and less coding means less room for bugs. I can't see any memory or CPU bound in such a simple function.

[cag8f]

Am I correct though that I will still have to hard-code each potential outcome into the database? i.e. I still have to enter into the database that a height of 75 and weight of 170 corresponds to group A, a height of 80 and a weight of 175 corresponds to group B, etc.

>> I can't see any memory or CPU bound in such a simple function

Are you saying that the SQLite method would be faster and/or use less memory than the Excel method?

edit: Also, a noob question--how do I reply to a single user in the SQLite mailing list? I see that people have responded to the question I posted to the mailing list. But when I click 'reply via email' at the bottom, I get an error page saying, "Requested archive does not exist."

Edited April 21, 2013 by cag8f

[kylomas]

cage8f,

Does this

My actual scenario involves upwards of 10 criteria (e.g. height, weight, age, etc), with upwards of 200 user-entered possibilities each.

mean that each row (person?) has a possibility of 10 characteristics and that there may be 200 rows (people) or that each criteria has a possiblity of 200 different values?

I'm not sure that we have enough definition to suggest a resonable solution regarding implementation.

kylomas

Edited April 21, 2013 by kylomas

[cag8f]

@kylomas: Each of the 10 different criteria has a potential of 200 potential values. A user would enter 1 of 200 potential heights, 1 of 200 potential weights, 1 of 200 potential ages, etc until the user did this or 10 different criteria. The program would then return the user's group.

[BrewManNH]

What determines the group they get put into? Because it sounds like a simple math problem at this point.

[cag8f]

The program would determine the group each user is assigned. The program would already know which height/weight combinations are assigned to which groups. It just waits for a user to enter his height/weight combination, then assigns him a group.

[BrewManNH]

That didn't answer the question I asked, what determines which group they get put into (a or ? Because it seems like the criteria would fall into ranges, and that could be done with something as simple as a Switch statement rather than making it so hard.

[cag8f]

The person writing the computer program would initially set all of the height/weight/group combinations. Then a user would enter his height and weight into the program, and the program would assign the user a group. The method the program used to do that is the reason for my post. Would a database solution be advantageous, keeping in mind that the actual scenario involves hundreds of thousands of potential combinations.

[BrewManNH]

Ok, you're not going to answer the question I asked I guess. I have asked you twice now what in your process will determine which group the person gets put into and twice you have repeated useless information. I am aware that the program will do it, what I don't know is how the program is supposed to figure out which group a person gets put into depending on the criteria you ask for. If you can answer that very simple question, we can probably proceed. If you won't/can't answer that, I will leave it up to others to help you.

[cag8f]

Sorry I'm a database noob here so I guess I'm not understanding your question correctly. I think your question is very similar (if not identical) to the question I'm asking in my original post. Yes I want my program to decide what group to assign each user. I know that one way to do this is to hard-code all of the potential height/weight/group combinations into an external data file (e.g. Excel), have the user enter his height/weight, have the program loop through all the potential height/weight/group combinations in that external data file, and output the user's group. I am wondering if a more advantageous method would use databases instead of looping over an external data file. Does that answer your question?

Edited April 21, 2013 by cag8f

[BrewManNH]

Not in the least.

You have 2 groups, A and B.

You have questions being asked to the users.

The answers to these questions somehow make them fall into one of the 2 groups

How do the group A people get assigned to group A and how do the group B people get assigned to group B after answering these questions.

There has to be some way to figure out how to assign them to one of the 2 groups by the questions asked, and the answers they give.

[cag8f]

Thanks for your patience but we don't seem to be getting very far.

>> There has to be some way to figure out how to assign them to one of the 2 groups by the questions asked, and the answers they give.

Here's the process:

1. External data file is created. This data file contains every single height/weight/group combination. Assume the computer program has decided each of these combinations based on his own whimsy.

2. Program asks 2 questions: "What is your height?" and "What is your weight?"

3. Program accesses this data file, searches for the line containing the user's height and weight, and returns the corresponding group.

Your issue is a little ambiguous to me--it it regarding step 1 or step 3?

[jchd]

Instead of using data tables, why don't you use a more universal algorithm like this page shows, using a simple of more sophisticated classification (age, ethnical criterion, etc)?

[BrewManNH]

Wouldn't it be easier to determine their group by some kind of algorithm that the height and weight fit into? In the way that a person's BMI is determined by the weight divided by the height, you don't need a chart or a database, you just need 2 numbers divided into each other. If the height/weight ratio is below a certain number they're in group A, if it's equal to or above a certain number they're in B.

I think you're WAY over-complicating it for what you're trying to do.

[cag8f]

Yes in this simple height/weight scenario, with only 2 criteria, a chart or database might be unnecessary. But I'm posing these questions because my actual scenario involves ~10 criteria, each with ~200 possibilities. My actual scenario also does not have an algorithm (e.g. body mass index) to determine each group. Would a database solution to such a scenario be ideal in terms of processing speed and memory usage?