Batch Invalid Number error?

Question

This is the strangest error I've ever encountered. While I've seen the:

Invalid number.  Numeric constants are either decimal (17),
hexadecimal (0x11), or octal (021).

error a million times, I can't understand what's wrong with this code:

if "%ss%" == "00" (
    set /a 3n%1+=%3
    set /a 3n%1+=%4
)

This gives:

C:\Users\...>if "00" == "00" (
set /a 3n2+=1
 set /a 3n2+=3
)
Invalid number.  Numeric constants are either decimal (17),
hexadecimal (0x11), or octal (021).

I can't figure out why this is happening.

Note: Even when the expression is false, the error still fires.

I don't think the code you posted can generate that error if the IF condition is false. That error message is only generated if the SET /A command is executed, which can never happen if the expression is false. The statement is always parsed, but it is only executed if the condition is true. — dbenham, Aug 31 '15 at 02:14

score 3 · Answer 1 · edited May 23 '17 at 11:52

3

The error comes from the fact that your variable name 3n2 starts with a number.

While the variable name is technically valid, starting it with a number is a really bad idea and you should never do it.

To avoid the error, change the first character of the variable name to a letter or an underscore.

if "%ss%" == "00" (
set /a _3n%1+=%3
set /a _3n%1+=%4
)

edited May 23 '17 at 11:52

Community

1
1

answered Aug 31 '15 at 00:51

SomethingDark

13,229
5
50
55

score 2 · Answer 2 · edited May 23 '17 at 10:26

I agree with SomethingDark's answer: the starting digit is causing the "wrong" behaviour and it can (must!) be solved changing the name of the variable.

The general batch parser identifies the set command, call the internal function asociated who sees the /A switch and delegates all the work to a different parser that handles the arithmetic syntax.

The source of the behaviour is the syntax allowed by the set /a arithmetic parser. While it seems obvious that set /a 3n2+=1 is adding one to a variable, the set /? help states that

The syntax for arithmetic set operations is set /a expression
The = is an operator

That is, while set /a can, and usually is used to, set the value of a variable, it allows things as set /a 1+1, without any variable involved. The variables are just one operand used by an assignation operator.

When the arithmetic parser starts its work, one of the first operations (after quote removal, initial space trim, ...) is to check if the next character to process in the input line is a digit, and if it is, the parser assumes that a numeric operand has been found and a conversion from the input command text to a number is needed to operate, and here the error is generated as 3n2 can not be converted to a number.

Of course, if the text can be converted to a number and no conversion error raises,

set /a 123=1

we will also find and error, a missing operator. The data at the left part of the equal sign has been identified as a number, and a number can not be placed as the left side of an assignation operator (a variable is needed), the parser expects a different operator that can not be found.

A similar (more obvious) case is

set /a =123+1

where we will get a missing operand error: the equal sign is a binary operator, and the left side operand can not be found.

Batch Invalid Number error?

2 Answers2