Open Nezrahm opened 2 months ago
Duplicate of #39
The number described in this original issue has higher precision than the maximum precision for a double (so some loss of precision is literally inevitable), but here are some numbers where JSON-Viewer loses precision unnecessarily (i.e., a better double-formatting algorithm could retain that precision):
[
-5.0975200963490517E-33,
-6.3809366960906694E-48,
-7.4034655373995265E-78,
-7.7377839245507245E-78,
1.2859736125485259E-22,
-3.5403485244736897E-88,
-6553693.3617752995,
-6.8141086401061905E-44,
5.5843284390697506E-71,
-4.3180528943019356E-77,
-5.8231387142089446E-34,
4.9686769279508796E-34,
-1.7271559814272259E-77,
-4.2596088775464085E-81,
3.7049941506206046E-71,
3.9742364030067576E-86,
3.1628542535929037E-89,
-268.52143589416397,
-2.2795405986944148E-93,
-7.3921473885461993E-95,
-5.1970028999668327E-45,
3.5937267475433058E-09,
7.2781313854297585E-75,
2.1823857766614118E-78,
8.9080568887277594E-11,
-898453.63759556494,
-6.5029793100887976E-55,
15650583.101212589,
-7.7306921203660293E-55,
4.3436060211811726E-74,
-7.1399486851522386E-90
]
The most recent version of JsonTools can reformat this file without losing precision on any of these numbers.
Let's take this simplified json as an example
{"MyDecimal":5.107091491635510056019771245}
If I format/pretty print this data then will the result be
This has effectively corrupted the data.