十进制值被 to_f 截断

Question

. Answers to this question are eligible for a +150 reputation bounty. Developer wants to to this question.

I have model called Item, where I am updating the unit_price (Data Type is Decimal) value, currently I am not putting any limit when storing the value, storing the value as it is. But now I can see this PG error PG::NumericValueOutOfRange , when the value exceeds the limit.

So I was just trying to limit the value and checking something in the console, Below is the data. (Here in the data I am not putting all the decimal values)

#<Item id: 167199, description: "192830139", category_id: 10327, unit_id: 5596, weight: 0.1e5, unit_price: 0.4083333333659917816764132553606237816656920077972709552126705653021442494641325536062378168e1

i = Item.find 167199

i.unit_price.to_f
=> 4.083333333659918

---

#<Item id: 167199, description: "192830139", category_id: 10327, unit_id: 5596, weight: 0.1e5, unit_price: 0.6511366980197836882065909262763993442019943880913510722934069011050182329156169820243980265070876781866034494363303661586489199452739290976143216266200531728395970406461889852558384421962422689303402903e-2

i.unit_price.to_f
=> 0.006511366980197837

Can I know what will be the reason the to_f automatically reduce the limit of the decimal? What will be best way to solve this issue, I was just thinking about some truncate with some limit.

Answer 1

Can I know what will be the reason the to_f automatically reduce the limit of the decimal?

The BigDecimal#to_f method will convert an arbitrary precision floating point decimal object into a standard double precision Float ing point number. Naturally, information will be lost during this conversion should the big decimal be more precise. This conversion can actually overflow or underflow if limits are exceeded.

I was just thinking about some truncate with some limit

There is a truncate(precision) method if you'd like explicit control over the precision of the result.

Answer 2

you can use ruby's built-in .truncate() method

for example:

floatNum = 1.222222222222222
truncatedNum = floatNum.truncate(3) #3 is the number of decimal places you want
puts floatNum #returns 1.222

another way is to use the .round() method

for example: