PHP is very popular for developing business applications. Often such applications needs to calculate numbers, for example when dealing with money.
Even if php is a high level programming language, dealing with something like money is error prone and often not proper implemented.
One of the reasons for this is that php doesn't have a Decimal
type. Often people use float's to do such calculations.
If you take a look at the floating point documentation there is a huge warning
So never trust floating number results to the last digit, and do not compare floating point numbers directly for equality.
It's quite easy to do this wrong and even if this just affect the last digit
this could result in a huge, hard to fix problem.
If you simply want to calculate a price, or do some basic math you probably don't want to think about Floating-Point Arithmetic at all.
The package comes with the classes Decimal
and Decimal[0-100]
the 0-100
indicates the precision.
This would allow you to write Something like (new Decimal2("1.99"))->add(new Decimal2("1.99"))
.
There are also helper functions to create Decimals: dec2("1.99")->add(dec2("1.99"))
. The Function dec2
returns a Decimal2
, dec99
could return an Instance of Decimal99
.
The package also comes with a Rational Type.
This could help you for example if you need to calculate something like 3 * (1/3)
. Using the rational type you could simply write
Rational::fromDecimal(dec0(3))->mul(new Rational(dec0(1), dec0(3))
. There is also a rat()
that looks a bit nicer.
you could represent 1 + 2
as 3
but for more complex numbers like the Rational
-Type this doesnt work well. For example you could not represent 1/3
as decimal without losing precision.
The Idea behind Lazy Calculation is, that instead of calculating numbers directly, just remember the calculation steps and calculate them as late as possible.
let's take a small look at this example:
$calculation = lazy_calc(dec(0))
->add(lazy_calc(rat(1, 3))->round(2))
->add(lazy_calc(rat(1, 3))->round(2))
->add(lazy_calc(rat(1, 3))->round(3));
static::assertSame(
'(((0 + round((1 / 3), scale = 2)) + round((1 / 3), scale = 2)) + round((1 / 3), scale = 3))',
$calculation->pretty()
);
As you can see here, $calculation
just remembers the complete calculation. You can now do different things with the calculation, for example just print it down or use different calculators to calculate a result.
It is also possible to add hints to numbers and calculation steps and pretty-print the calculation with annotations.
This makes it really easy to understand huge calculations. For example this could be very interesting if you have to calculate a basket with a lot of different items, rebates, conditions and so on. Debugging the basket would be a lot easier if you could simply dump all calculation steps, annotated and explained.