string pattaren-matching using awk

Question

(Couldn't come with any better title)

I'm trying to convert a number of lines, like these:

#define GENERIC_TYPE_METER_PULSE                     0x30 /*Pulse Meter*/
#define SPECIFIC_TYPE_NOT_USED                       0x00 /*Specific Device Class not used*/
......
#define MFG_ID_WAYNE_DALTON                          0x0008   //Wayne Dalton
#define MFG_ID_WILSHINE_HOLDING_CO_LTD               0x012D   //Wilshine Holding Co., Ltd
#define MFG_ID_WIDOM                                 0x0149   //wiDom
......
#define COMMAND_CLASS_ALARM                              0x71
#define COMMAND_CLASS_ALARM_V2                           0x71
#define COMMAND_CLASS_NOTIFICATION_V3                    0x71
#define COMMAND_CLASS_NOTIFICATION_V4                    0x71

in a file (c++) to something like these (for Java):

SPECIFIC_TYPE_NOT_USED((byte)0x00)             /*Specific Device Class not used*/
MFG_ID_WILSHINE_HOLDING_CO_LTD((byte)0x012D)   //Wilshine Holding Co., Ltd
COMMAND_CLASS_ALARM((byte)0x71)
....

I came up with this:

gawk '/^#define/ && / [[:xdigit:]]/ { printf "%s((byte)%s)\n",$2,$3 }'

but there are two issues - it doesn't work with awk; requires GNU-awk (gawk) and I also need the end-of-the-line comments in the output whenever available. How can I do that? I'm especially interested using awk but can live with sed as well. Cheers!!

Answer 1

one sed solution, i just tested once, so first test some cases

[[bash_prompt$]]$ cat log; echo "########";sed -e 's/#define \([^ ]*\)[ ]*\([^ ]*\)/\1((byte)\2)/g' log
#define GENERIC_TYPE_METER_PULSE                     0x30 /*Pulse Meter*/
#define SPECIFIC_TYPE_NOT_USED                       0x00 /*Specific Device Class not used*/
#define MFG_ID_WAYNE_DALTON                          0x0008   //Wayne Dalton
#define MFG_ID_WILSHINE_HOLDING_CO_LTD               0x012D   //Wilshine Holding Co., Ltd
#define MFG_ID_WIDOM                                 0x0149   //wiDom
#define COMMAND_CLASS_ALARM                              0x71
#define COMMAND_CLASS_ALARM_V2                           0x71
#define COMMAND_CLASS_NOTIFICATION_V3                    0x71
#define COMMAND_CLASS_NOTIFICATION_V4                    0x71
########
GENERIC_TYPE_METER_PULSE((byte)0x30) /*Pulse Meter*/
SPECIFIC_TYPE_NOT_USED((byte)0x00) /*Specific Device Class not used*/
MFG_ID_WAYNE_DALTON((byte)0x0008)   //Wayne Dalton
MFG_ID_WILSHINE_HOLDING_CO_LTD((byte)0x012D)   //Wilshine Holding Co., Ltd
MFG_ID_WIDOM((byte)0x0149)   //wiDom
COMMAND_CLASS_ALARM((byte)0x71)
COMMAND_CLASS_ALARM_V2((byte)0x71)
COMMAND_CLASS_NOTIFICATION_V3((byte)0x71)
COMMAND_CLASS_NOTIFICATION_V4((byte)0x71)
[[bash_prompt$]]$

for a proper format, ( in my sense, it is aligning the comment ) , i was not able to come up with any sed solution, ( i am interested in one ) here is an awk solution ( looks like a little clumsy, but self explanatory..

[[bash_prompt$]]$ awk '{str=sprintf("%s((byte)%s)", $2, $3);len=length(str); \
> for(i=50;i>len;i--) str=sprintf("%s ", str); print str (NF>3?substr($0,index($0,$4)):"")}' log
GENERIC_TYPE_METER_PULSE((byte)0x30)              /*Pulse Meter*/
SPECIFIC_TYPE_NOT_USED((byte)0x00)                /*Specific Device Class not used*/
MFG_ID_WAYNE_DALTON((byte)0x0008)                 //Wayne Dalton
MFG_ID_WILSHINE_HOLDING_CO_LTD((byte)0x012D)      //Wilshine Holding Co., Ltd
MFG_ID_WIDOM((byte)0x0149)                        //wiDom
COMMAND_CLASS_ALARM((byte)0x71)
COMMAND_CLASS_ALARM_V2((byte)0x71)
COMMAND_CLASS_NOTIFICATION_V3((byte)0x71)
COMMAND_CLASS_NOTIFICATION_V4((byte)0x71)

Answer 2

You can use this awk one-liner:

awk '$1=="#define" && $3~/[0-9]+$/{printf "%s((byte)%s)", $2, $3; $1=$2=$3=""; print}' file

EDIT: Here is an awk that you can try for better alignment of comments:

awk '$1 == "#define" && $3 ~ /[0-9]+$/{s=sprintf("%s((byte)%s)", $2, $3); $1=$2=$3="";
     printf("%-50s\t%s\n", s, $0)}'

OUTPUT:

GENERIC_TYPE_METER_PULSE((byte)0x30)                      /*Pulse Meter*/
SPECIFIC_TYPE_NOT_USED((byte)0x00)                        /*Specific Device Class not used*/
MFG_ID_WAYNE_DALTON((byte)0x0008)                         //Wayne Dalton
MFG_ID_WIDOM((byte)0x0149)                                //wiDom
COMMAND_CLASS_ALARM((byte)0x71)                           
COMMAND_CLASS_ALARM_V2((byte)0x71)                        
COMMAND_CLASS_NOTIFICATION_V3((byte)0x71)                 
COMMAND_CLASS_NOTIFICATION_V4((byte)0x71)      

Answer 3

$ cat tst.awk
/^#/ {
    hd[++nr] = sprintf("%s((byte)%s)", $2, $3)
    lgth = length(hd[nr])
    maxLgth = (lgth > maxLgth ? lgth : maxLgth)

    sub(/[^/]+/,"")
    tl[nr] = $0
}

END {
    for (i=1; i<=nr; i++)
        printf "%-*s%s\n", maxLgth+2, hd[i], tl[i]
}
$ awk -f tst.awk file
GENERIC_TYPE_METER_PULSE((byte)0x30)          /*Pulse Meter*/
SPECIFIC_TYPE_NOT_USED((byte)0x00)            /*Specific Device Class not used*/
MFG_ID_WAYNE_DALTON((byte)0x0008)             //Wayne Dalton
MFG_ID_WILSHINE_HOLDING_CO_LTD((byte)0x012D)  //Wilshine Holding Co., Ltd
MFG_ID_WIDOM((byte)0x0149)                    //wiDom
COMMAND_CLASS_ALARM((byte)0x71)
COMMAND_CLASS_ALARM_V2((byte)0x71)
COMMAND_CLASS_NOTIFICATION_V3((byte)0x71)
COMMAND_CLASS_NOTIFICATION_V4((byte)0x71)

Change the 2 in maxLgth+2 to whatever spacing you want between the longest value and it's associated comment.