rleyba828
asked on
Requesting assistance with simple bash script to convert text list to csv
Hi team,
just need help with a script that can convert a regular TXT file into a CSV file. I have an "expect" tcl program that logs in to multiple network devices and extract SSL certificate information from each. I then collect the output to a file. To make the output useful, I would have to convert the generated "certoutput.txt" file to a "certoutput.csv" so I can import it to excel and do further filtering and manipulation.
Here is how the raw output looks like. It can possibly be hundreds of lines long, and I just show here how the basic *repeating* format would typically look like. In the case below, there are 3 successful SSL records and 1 error record.
I just need the output to look like this:
*The idea is to ignore all lines where it will say "error" , or "unable to load certificate" or the actual lines with the prompt "[lnetadmin@load-balancer- A] ssl.crt # "
Thanks very much.
just need help with a script that can convert a regular TXT file into a CSV file. I have an "expect" tcl program that logs in to multiple network devices and extract SSL certificate information from each. I then collect the output to a file. To make the output useful, I would have to convert the generated "certoutput.txt" file to a "certoutput.csv" so I can import it to excel and do further filtering and manipulation.
Here is how the raw output looks like. It can possibly be hundreds of lines long, and I just show here how the basic *repeating* format would typically look like. In the case below, there are 3 successful SSL records and 1 error record.
[lnetadmin@load-balancer-A] ssl.crt #
openssl x509 -issuer -subject -startdate -enddate -in cert1.intranet.mycompany.com.crt -noout
issuer= /DC=US/DC=COM/DC=Mycompany/DC=USDOM/DC=infra/CN=someissuer
subject= /C=US/ST=California/L=mycity/O=myorg/OU=enterprise/CN=test1.bpm.intranet.mycompany.com
notBefore=Feb 20 03:58:08 2013 GMT
notAfter=Apr 16 03:58:08 2014 GMT
[lnetadmin@load-balancer-A] ssl.crt #
openssl x509 -issuer -subject -startdate -enddate -in CERT2.intranet.mycompany.com.crt -noout
issuer= /DC=US/DC=COM/DC=Mycompany/DC=USDOM/DC=infra/CN=someissuer
subject= /C=US/ST=California/L=mycity/O=myorg/OU=enterprise/CN=test2.bpm.intranet.mycompany.com
notBefore=Feb 20 03:58:08 2013 GMT
notAfter=Apr 16 03:58:08 2014 GMT
[lnetadmin@load-balancer-A] ssl.crt #
openssl x509 -issuer -subject -startdate -enddate -in allcerts.txt -noout
unable to load certificate
29648:error:0906D06C:PEM routines:PEM_read_bio:no start line:pem_lib.c:647:Expecting: TRUSTED CERTIFICATE
[lnetadmin@load-balancer-A] ssl.crt #
openssl x509 -issuer -subject -startdate -enddate -in CERT3.crt -noout
issuer= /DC=US/DC=COM/DC=Mycompany/DC=USDOM/CN=someissuer
subject= /DC=US/DC=COM/DC=Mycompany/DC=USDOM/DC=infra/CN=test3.intranet.mycompany.com
notBefore=Oct 29 08:31:41 2009 GMT
notAfter=Oct 29 08:41:41 2014 GMT
[lnetadmin@load-balancer-A] ssl.crt #
[netadmin@myserver]$
I just need the output to look like this:
CERTID, ISSUER, CNNAME, ISSUEDATE, EXPIRATION
cert1.intranet.mycompany.com.crt , DC=US/DC=COM/DC=Mycompany/DC=USDOM/DC=infra/CN=someissuer,test1.bpm.intranet.mycompany.com,Feb 20 03:58:08 2013,Apr 16 03:58:08 2014
CERT2.intranet.mycompany.com.crt,DC=US/DC=COM/DC=Mycompany/DC=USDOM/DC=infra/CN=someissuer,test2.bpm.intranet.mycompany.com, Feb 20 03:58:08 2013, Apr 16 03:58:08 2014
CERT3.crt -noout,DC=US/DC=COM/DC=Mycompany/DC=USDOM/CN=someissuer, test3.intranet.mycompany.com, Oct 29 08:31:41 2009,Oct 29 08:41:41 2014
*The idea is to ignore all lines where it will say "error" , or "unable to load certificate" or the actual lines with the prompt "[lnetadmin@load-balancer-
Thanks very much.
perl -l -0133ne '(1..1)&&print "CERTID, ISSUER, CNNAME, ISSUEDATE, EXPIRATION";print "$1 , $2,$3, $4" if /openssl.*?-in\s+(\S+).*?i ssuer=\s*( \S+).*?sub ject=.*?CN =(\S+).*?n otAfter=([ ^\n]*)/s'
SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Just saw you also want to get rid of "GMT":
awk ' BEGIN {print "CERTID, ISSUER, CNNAME, ISSUEDATE, EXPIRATION"}
{if($1~/^openssl/) {FS="-in | -noout"; S=$2}
if($1~/^issuer=/) {FS="issuer= "; sub("^/","",$2); printf "%s,%s,", S, $2}
if($1~/^subject=/) {FS="CN="; printf "%s,", $2}
if($1~/^notBefore/) {FS="="; sub(" GMT","",$2); printf "%s,", $2}
if($1~/^notAfter/) {FS="="; sub(" GMT","",$2); print $2}}' certoutput.txt > certoutput.csv
awk ' BEGIN {print "CERTID, ISSUER, CNNAME, ISSUEDATE, EXPIRATION"}
{if($1~/^openssl/) {FS="-in | -noout"; S=$2}
if($1~/^issuer=/) {FS="issuer= "; sub("^/","",$2); printf "%s,%s,", S, $2}
if($1~/^subject=/) {FS="CN="; printf "%s,", $2}
if($1~/^notBefore/) {FS="="; sub(" GMT","",$2); printf "%s,", $2}
if($1~/^notAfter/) {FS="="; sub(" GMT","",$2); print $2}}' certoutput.txt > certoutput.csv
ASKER
Hi Ozo,
Thanks for this, the output of your file using the example above, is now printing ok as:
However, on my real data set which contains 70 certificates, when I simply do a cat certoutput.csv, then the line endings look weird and the "notafter" field appears to overlap the first few characters of that line. However, when I do gedit certoutput.csv, everything seems to line up properly, except that after the "notbefore" field, the next comma and the "notafter" field is forced on the next line. This happened on every single record (70 certs in my case) of the certoutput.csv file. I thought this was just due to the screen wrap of my monitor, but when I opened on a very wide screen, the "notafter" field REALLY was being forced on the next line.
Unfortunately, I can't post the output here as they are all our customer's certificates and it is too difficult for me to sanitize the output for publishing on this forum.
Finally, could you kindly assist a non Perl expert like me by letting me know what "perl -l -0133ne '(1..1)&&" means? The next lines of your code seem to make sense as they are all regex constructs, but am not sure about the first part.
Also, I have to run the same script on about 92 other devices and I was going to do it this way.
and the content of this file would be
and the final output would contain the "device name" as the first column, like this....
would very much appreciate if you could just add the extra line or two needed in your code that would print out the device name as the first field in the output file. thanks .
for woolmilkporc, using the same example above, yours is coming out like this for some reason:
thanks very much.
Thanks for this, the output of your file using the example above, is now printing ok as:
[lnetadmin@myserver]$ cat certoutput.csv
CERTID, ISSUER, CNNAME, ISSUEDATE, EXPIRATION
cert1.intranet.mycompany.com.crt , /DC=US/DC=COM/DC=Mycompany/DC=USDOM/DC=infra/CN=someissuer,test1.bpm.intranet.mycompany.com, Feb 20 03:58:08 2013 GMT, Apr 16 03:58:08 2014 GMT
CERT2.intranet.mycompany.com.crt , /DC=US/DC=COM/DC=Mycompany/DC=USDOM/DC=infra/CN=someissuer,test2.bpm.intranet.mycompany.com, Feb 20 03:58:08 2013 GMT, Apr 16 03:58:08 2014 GMT
CERT3.crt , /DC=US/DC=COM/DC=Mycompany/DC=USDOM/CN=someissuer,test3.intranet.mycompany.com, Oct 29 08:31:41 2009 GMT, Oct 29 08:41:41 2014 GMT
However, on my real data set which contains 70 certificates, when I simply do a cat certoutput.csv, then the line endings look weird and the "notafter" field appears to overlap the first few characters of that line. However, when I do gedit certoutput.csv, everything seems to line up properly, except that after the "notbefore" field, the next comma and the "notafter" field is forced on the next line. This happened on every single record (70 certs in my case) of the certoutput.csv file. I thought this was just due to the screen wrap of my monitor, but when I opened on a very wide screen, the "notafter" field REALLY was being forced on the next line.
Unfortunately, I can't post the output here as they are all our customer's certificates and it is too difficult for me to sanitize the output for publishing on this forum.
Finally, could you kindly assist a non Perl expert like me by letting me know what "perl -l -0133ne '(1..1)&&" means? The next lines of your code seem to make sense as they are all regex constructs, but am not sure about the first part.
Also, I have to run the same script on about 92 other devices and I was going to do it this way.
extract-all-certs.sh
and the content of this file would be
perl -e yourscript device-1-certs.txt > device-1-certs.csv
perl -e yourscript device-2-certs.txt > device-2-certs.csv
perl -e yourscript device-3-certs.txt > device-3-certs.csv
perl -e yourscript device-4-certs.txt > device-4-certs.csv
perl -e yourscript device-5-certs.txt > device-5-certs.csv
........
perl -e yourscript device-92-certs.txt > device-92-certs.csv
and the final output would contain the "device name" as the first column, like this....
[netadmin@myserver]$ cat device-1-certs.csv
DEVICE-NAME,CERTID, ISSUER, CNNAME, ISSUEDATE, EXPIRATION
device-1,cert1.intranet.mycompany.com.crt , /DC=US/DC=COM/DC=Mycompany/DC=USDOM/DC=infra/CN=someissuer,test1.bpm.intranet.mycompany.com, Feb 20 03:58:08 2013 GMT, Apr 16 03:58:08 2014 GMT
device-1,CERT2.intranet.mycompany.com.crt , /DC=US/DC=COM/DC=Mycompany/DC=USDOM/DC=infra/CN=someissuer,test2.bpm.intranet.mycompany.com, Feb 20 03:58:08 2013 GMT, Apr 16 03:58:08 2014 GMT
device-1,CERT3.crt , /DC=US/DC=COM/DC=Mycompany/DC=USDOM/CN=someissuer,test3.intranet.mycompany.com, Oct 29 08:31:41 2009 GMT, Oct 29 08:41:41 2014 GMT
would very much appreciate if you could just add the extra line or two needed in your code that would print out the device name as the first field in the output file. thanks .
for woolmilkporc, using the same example above, yours is coming out like this for some reason:
[lnetadmin@myserver]$ cat certoutput.csv
CERTID, ISSUER, CNNAME, ISSUEDATE, EXPIRATION
x509,,,,Apr 16 03:58:08 2014
,,,,Apr 16 03:58:08 2014
CERT3.crt,,,,Oct 29 08:41:41 2014
[lnetadmin@myserver]$
thanks very much.
Are you on Solaris?
If so, use /usr/xpg4/bin/awk or "nawk".
Besides that, when run against the data you posted in the Q my version runs just fine under Linux and AIX.
If so, use /usr/xpg4/bin/awk or "nawk".
Besides that, when run against the data you posted in the Q my version runs just fine under Linux and AIX.
ASKER
Hi Ozo,
I did more analysis, and in my real data set, after the issuedate field in all the records, there is a [CR] character, that's why it was forcing the next field to start on the next line (in case when I open the file in gedit) or on the SAME line, as is the case when I do a simple cat certoutput.csv. How do we remove this extraneous [CR] character?
Thanks.
I did more analysis, and in my real data set, after the issuedate field in all the records, there is a [CR] character, that's why it was forcing the next field to start on the next line (in case when I open the file in gedit) or on the SAME line, as is the case when I do a simple cat certoutput.csv. How do we remove this extraneous [CR] character?
Thanks.
s/\r//g
ASKER
Hi woolmilkporc, not sure what was going on , I am using on plain old Linux OpenClient 2.6.32-358.11
For Ozo, thanks for the search and replace string...that's what I need.
Thanks.
For Ozo, thanks for the search and replace string...that's what I need.
Thanks.