Setting input parameter to be float for Python script tool?

Question

I made a Python (2.7.8) script, where one of the parameters is height (float) and I wan to make tool in ArcGIS 10.3 now. However in data type list there is only double. If I input integer value into this parameter field (e.g. -480), the tool works fine. If i input e.g. - 480,5 i get error

Traceback (most recent call last): File "D:\HETMAN\GIS\Projekt Model Złoża\Sktypy.py", line 27, in ValueError: invalid literal for float(): -480,1

Failed to execute (SSKv1).

This is a fragment of code, where i try to convert value to float:

kontakt_dbl = sys.argv[3]
kontakt = float(kontakt_dbl)

How to convert double value into float and not to get error?

Answer 1

Your issue is arising due to the use of a comma as the decimal mark. It may help to understand that in your script you are not converting a double into a float, you are converting a string into a float. The string, when a comma is used as the decimal mark, is invalid.

I would make use of the Replace method here to ensure that a decimal point is used in the script:

kontakt_dbl = sys.argv[3]
kontakt = float(kontakt_dbl.replace(",","."))

In this case if a user inputs "-480,5", it is changed to "-480.5" which is a valid input.

Regarding data types, the data type Double is a floating point number. ArcGIS uses two types of floating point numbers, differentiated by precision: Float (single-precision) and Double (double-precision). The following page explains the difference in more detail: http://resources.arcgis.com/EN/HELP/MAIN/10.1/index.html#/ArcGIS_field_data_types/005s0000000p000000/

• Cyrbil: You may want to use locale.atof() to let the parsing of float depend of the environment locale.
• Cyrbil: @Luke: see answer below.
• GeoStoneMarten: Dont use locale because it depend of production process data. If this is not the same software or a same language you have got a problem because number can be stored with point or comma... In this condition, i prefer cast data because i know destination format but not source format. If it's necessary, I check with "if isinstance( v, basestring): "

Answer 2

You can use the locale module to let the parsing of float depends of the environment locale.

>>> import locale
>>> locale.setlocale(locale.LC_ALL, '') # (optional here) use user's preferred locale
>>> locale.getlocale()
('en_GB', 'ISO8859-1') # my current env locale
>>> locale.atof('12,3')
123.0 # wrongly parsed
# with another locale, parsing is correct
>>> locale.setlocale(locale.LC_ALL, 'de_DE')
('de_DE', 'ISO8859-1')
>>> locale.atof('12,3')
12.3

Information about a local can be seen with:

>>> locale.localeconv()
{
 'mon_decimal_point': ',',
 'int_frac_digits': 2,
 'p_sep_by_space': 1,
 'frac_digits': 2,
 'thousands_sep': '.',
 'n_sign_posn': 1,
 'decimal_point': ',',
 'int_curr_symbol': 'EUR ',
 'n_cs_precedes': 0,
 'p_sign_posn': 1,
 'mon_thousands_sep': '.',
 'negative_sign': '-',
 'currency_symbol': 'EUR ',
 'n_sep_by_space': 1,
 'mon_grouping': [3, 0],
 'p_cs_precedes': 0,
 'positive_sign': '',
 'grouping': [3, 0]
}