fix(numberFilter): format large numbers correctly

[gauravaror]

format large number to correct form instead of NAN.00 currently.
in previous code when first time we add the exponent with fractionSize
(number.toString()+'e'+fractionsize)
and then convert it to an integer using ,
+(number.toString()+'e'+fractionsize)
when the number is very large and number will be represented by using exponent. for example 12345868059685210000 will be represented by 1.234586805968521e+21. as getting string in exponent is not handles, we get the number as NAN.
Handle when the number is being represented in exponent form, by adjusting exponent of number with fraction size:

Closes #8674

[lgalfaso]

LGTM

[caitp]

LGTM with nit:

This routine is very cryptically named, and it's pretty hard to follow the long expression on the third line --- please organize this better so that it's easier to understand, and add some comments --- some additional test cases would also be helpful just to make sure you don't surprise yourself in some cases.

I don't think the name timesPow10 is particularly helpful here either, should give this a more meaningful name

[gauravaror]

thanks for the valuable inputs, will do the changes you suggested.

i thought of name of getMultipliedNumber, but its not a very meaningful name either and i am not able to come-up with a better name. any suggestions to help me out

[caitp]

addScientificNotation?

[gauravaror]

Its not exactly doing that, this funcion changes the number by dividing or multiplying it with e(fraction) and if the number is smaller than e32 then scientific notation is not returned by function , normal number is returned.. so this name isn't very meaningful to the function

[caitp]

we don't really need to talk about the multiplying or dividing aspect, because it's not important. What we're doing is modifying the number by adding (or modifying) the E notation, and converting the result to a number --- yes this has the effect of giving us a different value, but it's a lot more specific to what it's actually doing

[gauravaror]

Basically i agree we should not talk about exponent, division and multiplication in function name because that the detail how its done not what its done(abstraction)
What this function does in all is shifting the decimal place on the number.

so shiftDecimalPlace?

[lgalfaso]

I am torn about this. Or we accept that there are some limits on number formatting or we go all the way to string manipulation (lgalfaso@2e980d9)

I think that the first option is best, but we do need to handle those limits gracefully

@caitp WDYT?

[gauravaror]

@lgalfaso most concerning limitation i think is of having precision only till 16th digit. are you pointing to handle this limitation or handling round-off ?

since we have only changed implementation if number is NAN. i.e its expecting to output with very high decimal place. Since number of digits are more than decimal place we want to see. rounding-off should not be a issue currently.

What other things need to be taken care in this ?

[lgalfaso]

@samuelharden as long as we do not end up with numbers as NAN.000000000000000000000000000000000000000 or something that does not make any sense, then I am ok

[gauravaror]

I think with these changes it won't end up with the NAN.00000 , but i will still try to think if there is any other erratic case where number might end up there.