Actually the names with consistent three decimal digits are better, because they:
- are easier to scan when printed in a list.
- are easier to parse (consistent formats are always better).
- sort correctly into numeric order using a standard sort. Here is an example showing that your new format does not sort correctly:
>> old = {'data_70.000_32.125.csv'; 'data_70.124_32.120.csv'; 'data_70.120_32.123.csv'};
>> sort(old) % correct numeric order
ans =
'data_70.000_32.125.csv'
'data_70.120_32.123.csv'
'data_70.124_32.120.csv'
>> new = {'data_70.0_32.125.csv'; 'data_70.124_32.120.csv'; 'data_70.12_32.123.csv'};
>> sort(new) % your format gives the wrong numeric order.
ans =
'data_70.0_32.125.csv'
'data_70.124_32.120.csv'
'data_70.12_32.123.csv'
Of course you can then use my FEX submission natsortfiles to get them back into the correct order, but why bother when those trailing zeros do the job anyway.
I don't see any good reason to remove those zeros, just disadvantages.
