Large numbers are dropped when importing

[aaronindm]

Probably a very easy answer, buy I am importing data where the highlighted field values are mostly 99999999, so Double 8 is fine.  The problem is that much farther down in the table, there are values higher than that - and those get set to 0.  Trying to change the SIZE in a Select Tool but the number is grayed out.  Is a Select Tool the right way to set a larger size?

[benakesh · Accepted solution]

Hi @aaronindm ,

Double can store very large numbers  : 1.7 x 10308 

Double

A standard double precision floating point value. It uses 8 bytes, and can represent values from +/- 1.7 x 10-308 to 1.7 x 10308 with 15 digits precision. A double uses a decimal that can be placed in any position. A double uses twice as many bits as a float and is generally used as the default data type for decimal values.

+/- 1.7 x 10-308 to 1.7 x 10308 with 15 digits precision

https://help.alteryx.com/current/Reference/DataFieldType.htm

 You can change  string size  . The numeric size is fixed  and it  depends on type  . Ex  int16  : 2  bytes  ,   int64: 8 bytes etc   

[CharlieS]

---

@benakesh wrote:

Hi @aaronindm ,

Double can store very large numbers  : 1.7 x 10308 

---

Keep in mind that's 1.7976931348623157 * 10^308, which is super huge. 

The size of the Double field is a representation of the bit storage. Doubles, by definition, will use 8 as compared to the 4 of a Float field. These "size" values are representative of both size and accuracy that these field types can store.

https://en.wikipedia.org/wiki/Double-precision_floating-point_format

If you have large, non-decimal numbers, perhaps an Int64 would be better? Max value is (2^63)-1 or 9,223,372,036,854,775,807 and does not sacrifice accuracy. 

[danilang]

Hi @aaronindm 

The discussions from @benakesh and @CharlieS are relevant, but your underlying issue seems to be that the input tool is forcing your data into a double field when an int64 or fixed decimal might be a better choice.  Where is your data coming from? Can you provide a sample?

Dan

[32bit]

A large number going into a double data type wouldn't set to 0, it would lose precision. I think probably you are facing a conversion error. Can you post a sample of the raw data that sets to 0? Also, what format is this? Excel, csv? Does is have symbols, separators, etc? Lots of unanswered questions.