Printable View
i have developed a program.lot of text and numeric inputs.
input needs to be validated.
1) All the validation is done right after we read it in by cinn>> ?
2) There might be lot of numeric data to read. and if we each time type out the validateion for the string,then NO code-reuse. .if we create a validation class,and have validate Int,validate string,will that be correct?wrong?
whats the best to design a program which requires lot of numeric and string input and needs to be validated.
Normally, it's best to validate input as soon as it has been read. After the first wrong piece of input, there are likely to be further errors. Continuing after bad input is not only a waste of time, but possibly dangerous.Quote:
Originally Posted by rahulsk1947
Validation usually is more about just the type of data. Sure, the input is an integer, but maybe it needs to be in the range of 0..10. A string might have to be all lower case. Etc. As the input constraints become more complicated, you are no longer doing just validation, but full-on parsing.Quote:
2) There might be lot of numeric data to read. and if we each time type out the validateion for the string,then NO code-reuse. .if we create a validation class,and have validate Int,validate string,will that be correct?wrong?
It completely depends on your requirements, which haven't been specified.Quote:
whats the best to design a program which requires lot of numeric and string input and needs to be validated.
Just refactor as you go. If you have one place that does what you want to do in a second or third place, then make it its own function at that time.
Any repeated four lines of code r more etc will do, just make it into a function and reuse it.
cin >> var will only read valid data for the datatype. It will flag an error and optionally throw an exception if the input data does not match what is expected of it.
Beyond that, if you find yourself doing the same kind of parsing for different input lines you should write a function.
It can also be very useful to use tables (perhaps using templates) to indicate various things. For example, an integer or a floating point value may have a min/max value. Strings may have a set of flags saying things like "must not have spaces", "only alpha chars", "Must be uppercase", "autoconvert to lowercase", and such things.
This way, you don't have to write lots of different functions to do certain things.
You may also want to write a function that takes a variable (string, integer/float) to be filled in, a set of flags or (min,max) values and a call-back function for validation.
A call-back function will solve things like "this field must be a US state" (so checking that the value is one of "AZ", "IA", "TX", "CA" etc), whilst the automatic part will take care of the "convert to upper case".
--
Mats