Antlr parsing conflict with tokens and identifiers - antlr3

Question

i am attempting to parse freeform strings (ANTLR 3) and am running into string/token issues. Googling has not yet helped me. Here is my grammar:
grammar TestHeader;
options {
language = Java;
output = AST;
}
tokens {
STOKEN_1 = 'SENDTO';
STOKEN_2 = 'SRCSYS';
STOKEN_3 = 'SOFTERROR';
}
#lexer::header {
package com.sample.parser;
}
#parser::header {
package com.sample.parser;
}
fragment TAG : /* empty rule: only used to change the 'type' */;
fragment CR: '\r';
fragment LF: '\n';
fragment SLASH: '/';
fragment HASH: '#';
fragment DIGIT: '0'..'9';
fragment SEMI_COLON: ';';
COLON: ':';
COMMA: ',';
DOT: '.';
LETTER: 'A'..'Z' | 'a'..'z';
HYPHEN: '-';
BRANCH_CHARS: '$' | '#' | HASH;
SPACE: (' ');
EOL: (CR? LF) | CR | SEMI_COLON;
DATE: ('0'..'1')? DIGIT SLASH ('0'..'3')? DIGIT SLASH DIGIT DIGIT;
TIME: ('0'..'2')? DIGIT COLON ('0'..'5')? DIGIT (COLON '0'..'5' DIGIT)?;
DECIMAL
:
DIGIT* ('.' DIGIT*)
{
if ($text.contains("/")) {
$type = DATE;
} else if ($text.contains(":")) {
$type = TIME;
}
}
;
NUMBER
:
'0'..'9' DIGIT*
{
if ($text.contains("/")) {
$type = DATE;
} else if ($text.contains(":")) {
$type = TIME;
}
}
;
SINGLELINE_COMMENT
:
'//-' ~('\r' | '\n')*
;
header
:
(word SPACE+ sequenceNumber SPACE+ type) EOL+
(SPACE* trailer) EOL+
;
word
:
(data += ~(SPACE))+
;
sequenceNumber
:
nn=NUMBER
;
type
:
LETTER
;
trailer
:
// (tt += ~(EOL ))*
SINGLELINE_COMMENT
;
With a test input
SY 260 O
//-$CJ******1
i get the following
0 null
-- 16 S
-- 16 Y
-- 22
-- 18 260
-- 22
-- 16 O
-- 13
-- 20 //-$CJ******1
-- 13
However with the following input
SR 260 O
//-$CJ******1
I get
line 1:2 mismatched character ' ' expecting 'C'
line 1:7 mismatched input 'O' expecting NUMBER
0 260 O
//-$CJ******1
Where am i going wrong? Any tips/help to get past this issues?