0

I have a raster with 32 bit unsigned values that I want to convert to a polygon shapefile. I ran the raster calculator in ArcGIS to remove the decimals by multiplying by a constant and then running raster calculator again but this time converting to integer.

The output integer values do not match what the values were before converting to an integer. I'm not sure what is happening but by doing it this way will not work especially if its coming up with totally different values.

Clarification

I am running ArcGIS 10.4.1

For example - the raster holds values from 7881.530000 to 53613.300000 Multiplying the raster values by 1000000 because of the 6 decimal places

Converting the raster to integer results in completely different values

Real Example:

Pixel value: 39961.472656

Multiply by Constant: 39961473024.000000 (this number is already different)

Converted to Integer: 1306767360

If the pixel values are much smaller, the results are fine

Example:

Pixel Value: 0.087640

Multiply by Constant: 87640.140625

Converted to Integer: 87640

The end result is to convert to a polygon shapefile where I will add a field and calculate the value to be based on the Integer Value where I divide it by the constant to get back to the original pixel value.

Improve this question
2
  • 1
    You've left out most of the details that would make this an answerable question: Version of ArcGIS, range of values before multiplying by a constant, the constant value, the range of values after multiplication, the details of the conversion process, the range of values after conversion, the expected range of values... Please Edit the question.
    – Vince
    Commented Dec 3, 2020 at 20:26
  • 3
    4-byte signed integer has a range of -2147483648 to +2147483647. If you multiply every value by 40 billion, they'll all overflow. Even one million will cause forty thousand to overflow (by an order of magnitude). The maximum constant you can use is 40,055 (signed int) or (80110 unsigned int). Frame challenge: Multiplying to get two billion discrete integer values is not going to result in an effective vector layer conversion.
    – Vince
    Commented Dec 3, 2020 at 22:22

0

Reset to default

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.