On Mon, Feb 22, 2016 at 11:51 PM, BartC <b...@freeuk.com> wrote:
> Yes, I mentioned that point. In the OP's language, the variables can be
> declared as a particular type; in Python they could perhaps be 'declared' by
> first assigning with 0, 0.0 or "" for example. But that would need reference
> parameters to make use of tidily.
What you could do is something like this:
def read_values(file, values):
"""Read one line from file and parse it into the given list.
Prepopulate values with a series of types or instances of types.
The line of text read will be parsed accordingly.
Returns the number of elements successfully read.
"""
line = file.readline()
for idx, val in enumerate(values):
if not isinstance(val, type):
val = type(val)
# Now add the logic to read an int, a float,
# a str, a Fraction, etc etc etc, from the line,
# breaking out of the loop if one can't be
# read, and stashing the result into values[idx]
# if it can.
return idx
with open("inputfile") as f:
values = [int, int, float, Fraction, str]
while read_values(f, values) == 5:
[index, level, percent, ratio, message] = values
This lets you "declare" the values' types, and use an input/output
parameter. It's not exactly the most Pythonic of techniques, but it
does kinda work. I suspect, though, that what you'd _really_ want is
something more like a sscanf string, or a "typed regex", which would
allow you to specify a bit more flexibly what you're looking for. You
could easily make a template string that has magic markers for "take
an integer", which would insert [0-9]+ (or something more elaborate)
and then capture it and call int() on the string before returning it.
ChrisA
--
https://mail.python.org/mailman/listinfo/python-list
