I use CR XI and the data source is a UNIX log file that has about 200,000 lines. Sample data is as follows:
I inserted a DUMMY blank line after every line in the code shown here to properly identify the original text fields (since due to the number of characters, the lines were wrapped in the Preview screen). In the source file, they are not there.
Out of these lines, all I am interested is in parsing the file for the text <errorcode> and use that line and the two previous lines concatenated in the original order. For example, the first occurrence of this will become as follows. (The log file somehow seems to insert carriage return where it should not be there).
I have CR text driver defined to read lines that are 2,000 characters long in FIXED format while reading this file.
For want of any other ideas, I started as follows:
@concat:
if left({error35_log.Title},9) = "<!DOCTYPE" then
previous({error35_log.Title}) & {error35_log.Title} & next({error35_log.Title})
From here, I use another formulao to identify the lines with errorcode in them (the lines that I really want for use in the report)
@errorcode_data:
if instr({@concat},"<ErrorCode>",1) <> 0
then
{@concat}
else
"NODATA" //NODATA is used to suppress the unwanted lines in Details section
After this step, I have different formulas to extract the information I need to report:
@Date
left({@errorcode_data},10)
@errorcode_string:
((extractstring({@errorcode_data},"<ErrorCode>",
"</ErrorCode>"
))
@TXN_string:
(extractstring({@errorcode_data},"<TxNumber>","</TxNumber>"
)
I tried using tonumber(((extractstring({@errorcode_data},"<ErrorCode>",
"</ErrorCode>")) in the above formula and I get error 'non-numeric values'
but when I use tonumber(@errorcode_string) in a separate formula, it works fine. Any ideas?
I am sure there are more straightforward methods to achieve the above and I can use any ideas / help in doing that.
I thought Record selection formula would be the best but I am not sure if I can combine the three lines in one while doing that. That will make the report work faster as I normally only have about 1% of the total lines have the error data.
Also, I want to group the errorcode and TXNumber field, but I do not get an option to use the formula in a group (only my original field is available for grouping or summary). I would like to know how I can achieve that.
I would appreciate any help / guidance I can get on this.
Thank you very much to all of you for your support.
TR
Code:
2007.01.09 13:41:30 [ExecuteThread: '44' for queue: 'weblogic.kernel.Default'] [5] weblogic APIEJB.PROCESSMESSAGE(String,int) Client request = <?xml version="1.0" encoding="UTF-8"?> <!DOCTYPE GetUserRequest SYSTEM "GetUserRequest.dtd"><GetUserRequest><CommonRequest CustomerSubType="FSM" CustomerType="FSM"><TxNumber>802</TxNumber><TxUUID>000g0h37
69</TxUUID><CusReferenceID>0999</CusReferenceID></
CommonRequest><RequestData><UUID>000g0h3769</UUID></
RequestData></GetUserRequest>
2007.01.09 13:41:30 [ExecuteThread: '42' for queue: 'weblogic.kernel.Default'] [5] weblogic APIEJB.PROCESSMESSAGE(String,int) Client request = <?xml version="1.0" encoding="UTF-8"?> <!DOCTYPE GetUserRequest SYSTEM "GetUserRequest.dtd"><GetUserRequest><CommonRequest CustomerSubType="FSM" CustomerType="FSM"><TxNumber>802</TxNumber><TxUUID>00g00469
cc</TxUUID><CusReferenceID>0999</CusReferenceID></
CommonRequest><RequestData><UUID>00g00469cc</UUID></
RequestData></GetUserRequest>
2007.01.09 13:41:32 [ExecuteThread: '46' for queue: 'weblogic.kernel.Default'] [5] weblogic APIEJB.PROCESSMESSAGE(String,int) Client request = <?xml version="1.0" encoding="UTF-8"?> <!DOCTYPE GetCountryProhibitsRequest SYSTEM "GetCountryProhibitsRequest.dtd"><GetCountryProhibitsRequest>
<CommonRequest CustomerSubType="WEBSITE" CustomerType="GTM"><TxNumber>832</TxNumber><TxUUID>0gggg015
dc</TxUUID><CusReferenceID>0999</CusReferenceID></
CommonRequest><RequestData><ExportCountryCode>US</
ExportCountryCode><ImportCountryCode>BS</ImportCountryCode>
<ShipDate>20070109</ShipDate></RequestData></
GetCountryProhibitsRequest>
2007.01.09 13:41:34 [ExecuteThread: '31' for queue: 'weblogic.kernel.Default'] [5] weblogic APIEJB.PROCESSMESSAGE(String,int) BACKEND response sending to Servlet = <?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE EditCommodityCountryReply SYSTEM "EditCommodityCountryReply.dtd">
<EditCommodityCountryReply><CommonReply><TxNumber>963</
TxNumber><TxUUID>00g00kafac</TxUUID></CommonReply><Error>
<ErrorCode>77003</ErrorCode><ErrorMsg/></Error></
EditCommodityCountryReply>
2007.01.09 13:41:59 [ExecuteThread: '31' for queue: 'weblogic.kernel.Default'] [5] weblogic OBJECTBROKER.PROCESSMESSAGE Invoking method TradeXpress.CustomerProfile.CustomerProfileEJB_jruilg_
ELOImpl.retrieveCommodityRecords
2007.01.09 13:41:59 [ExecuteThread: '31' for queue: 'weblogic.kernel.Default'] [5] weblogic OBJECTBROKER.PROCESSMESSAGE TX= 851 Invoke time= 7
2007.01.09 13:42:00 [ExecuteThread: '31' for queue: 'weblogic.kernel.Default'] [5] weblogic APIEJB.PROCESSMESSAGE(String,int) BACKEND response sending to Servlet = <?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE DetermineDocsForCommodityReply SYSTEM "DetermineDocsForCommodityReply.dtd">
<DetermineDocsForCommodityReply><CommonReply><TxNumber>
920</TxNumber><TxUUID>g0g00o4f17</TxUUID></CommonReply>
<Error><ErrorCode>55002</ErrorCode><ErrorMsg>[854]</
ErrorMsg></Error></DetermineDocsForCommodityReply>
2007.01.09 13:42:00 [ExecuteThread: '31' for queue: 'weblogic.kernel.Default'] [5] weblogic OBJECTBROKER.PROCESSMESSAGE Invoking method TradeXpress.DocDetermination.DocDeterminationEJB_jisdhc_
ELOImpl.determineDocsforCommodity
2007.01.09 13:42:01 [ExecuteThread: '46' for queue: 'weblogic.kernel.Default'] [5] weblogic APIEJB.PROCESSMESSAGE(String,int) Client request = <?xml version="1.0" encoding="UTF-8"?> <!DOCTYPE GetUserRequest SYSTEM "GetUserRequest.dtd"><GetUserRequest><CommonRequest CustomerSubType="FSM" CustomerType="FSM"><TxNumber>802</TxNumber><TxUUID>000g01e4
00</TxUUID><CusReferenceID>0999</CusReferenceID></
CommonRequest><RequestData><UUID>000g01e400</UUID></
RequestData></GetUserRequest>I inserted a DUMMY blank line after every line in the code shown here to properly identify the original text fields (since due to the number of characters, the lines were wrapped in the Preview screen). In the source file, they are not there.
Out of these lines, all I am interested is in parsing the file for the text <errorcode> and use that line and the two previous lines concatenated in the original order. For example, the first occurrence of this will become as follows. (The log file somehow seems to insert carriage return where it should not be there).
Code:
2007.01.09 13:41:34 [ExecuteThread: '31' for queue: 'weblogic.kernel.Default'] [5] weblogic APIEJB.PROCESSMESSAGE(String,int) BACKEND response sending to Servlet = <?xml version="1.0" encoding="UTF-8"?><!DOCTYPE EditCommodityCountryReply SYSTEM "EditCommodityCountryReply.dtd"><EditCommodityCountryReply>
<CommonReply><TxNumber>963</TxNumber><TxUUID>00g00kafac
</TxUUID></CommonReply><Error><ErrorCode>77003</ErrorCode>
<ErrorMsg/></Error></EditCommodityCountryReply>I have CR text driver defined to read lines that are 2,000 characters long in FIXED format while reading this file.
For want of any other ideas, I started as follows:
@concat:
if left({error35_log.Title},9) = "<!DOCTYPE" then
previous({error35_log.Title}) & {error35_log.Title} & next({error35_log.Title})
From here, I use another formulao to identify the lines with errorcode in them (the lines that I really want for use in the report)
@errorcode_data:
if instr({@concat},"<ErrorCode>",1) <> 0
then
{@concat}
else
"NODATA" //NODATA is used to suppress the unwanted lines in Details section
After this step, I have different formulas to extract the information I need to report:
@Date
left({@errorcode_data},10)
@errorcode_string:
((extractstring({@errorcode_data},"<ErrorCode>",
"</ErrorCode>"
))@TXN_string:
(extractstring({@errorcode_data},"<TxNumber>","</TxNumber>"
)
I tried using tonumber(((extractstring({@errorcode_data},"<ErrorCode>",
"</ErrorCode>"
)) in the above formula and I get error 'non-numeric values'but when I use tonumber(@errorcode_string) in a separate formula, it works fine. Any ideas?
I am sure there are more straightforward methods to achieve the above and I can use any ideas / help in doing that.
I thought Record selection formula would be the best but I am not sure if I can combine the three lines in one while doing that. That will make the report work faster as I normally only have about 1% of the total lines have the error data.
Also, I want to group the errorcode and TXNumber field, but I do not get an option to use the formula in a group (only my original field is available for grouping or summary). I would like to know how I can achieve that.
I would appreciate any help / guidance I can get on this.
Thank you very much to all of you for your support.
TR