https://bz.apache.org/bugzilla/show_bug.cgi?id=63614
--- Comment #6 from [email protected] --- The Command Line I Used is "C:\apache-jmeter-5.1.1\bin\jmeter.bat" --nongui --remotestart 10.89.8.23,10.89.8.19,10.89.8.21,10.89.8.22,10.89.8.20 --testfile C:\JMeterLoadTests\SDIS_REST\SDIS_AND_QRS_REST_Sampler.jmx --logfile "C:\JMeterLoadTests\SDIS_REST\D07292019_110334\SDIS\log\20KB\SDIS.log" --reportatendofloadtests --reportoutputfolder "C:\JMeterLoadTests\SDIS_REST\D07292019_110334\SDIS\report\20KB\SDIS_75vusers" --remoteexit @@@@@@@@@@@@@@@@@@@ I am copying code chunk taken from the class of org.apache.jmeter.save.CSVSaveService of Source of Jmeter master branch(from GIT code repo). I beleive this code is unable to parse the entries of errored samples correctly. This code parses each line into a String Array. For example, if you compare a successful entry and a failed entry: Failed Entry from the report file: 2019/07/29 17:25:46.807,21034,ECMaaSRESTCreateDocSampler,500,Request Failed,10.89.8.21-Thread Group 1-2,,false,Test failed: text expected to contain /Successfully performed action/,14,0,15,15,null,0,0,0 Success Entry from report file: 2019/07/29 17:26:07.523,303,ECMaaSRESTCreateDocSampler,200,Successfully performed action: HOST_IPADDRESS =10.89.8.21,10.89.8.21-Thread Group 1-13,,true,,2539,0,15,15,null,0,0,0 The above both entries have same number of tokens. But the Jmeter code is wrongly tokenizing the failed entry. The column count and token count becoming unequal. @@@@@@@@@@@@@@@@ code from line # 988 to 1075 @@@@@@@@@@@@@@@ public static String[] csvReadFile(BufferedReader infile, char delim) throws IOException { int ch; ParserState state = ParserState.INITIAL; List<String> list = new ArrayList<>(); CharArrayWriter baos = new CharArrayWriter(200); boolean push = false; while (-1 != (ch = infile.read())) { push = false; switch (state) { case INITIAL: if (ch == QUOTING_CHAR) { state = ParserState.QUOTED; } else if (isDelimOrEOL(delim, ch)) { push = true; } else { baos.write(ch); state = ParserState.PLAIN; } break; case PLAIN: if (ch == QUOTING_CHAR) { baos.write(ch); throw new IOException( "Cannot have quote-char in plain field:[" + baos.toString() + "]"); } else if (isDelimOrEOL(delim, ch)) { push = true; state = ParserState.INITIAL; } else { baos.write(ch); } break; case QUOTED: if (ch == QUOTING_CHAR) { state = ParserState.EMBEDDEDQUOTE; } else { baos.write(ch); } break; case EMBEDDEDQUOTE: if (ch == QUOTING_CHAR) { baos.write(QUOTING_CHAR); // doubled quote => quote state = ParserState.QUOTED; } else if (isDelimOrEOL(delim, ch)) { push = true; state = ParserState.INITIAL; } else { baos.write(QUOTING_CHAR); throw new IOException( "Cannot have single quote-char in quoted field:[" + baos.toString() + "]"); } break; default: throw new IllegalStateException("Unexpected state " + state); } // switch(state) if (push) { if (ch == '\r') {// Remove following \n if present infile.mark(1); if (infile.read() != '\n') { infile.reset(); // did not find \n, put the character // back } } String s = baos.toString(); list.add(s); baos.reset(); } if ((ch == '\n' || ch == '\r') && state != ParserState.QUOTED) { break; } } // while not EOF if (ch == -1) {// EOF (or end of string) so collect any remaining data if (state == ParserState.QUOTED) { throw new IOException("Missing trailing quote-char in quoted field:[\"" + baos.toString() + "]"); } // Do we have some data, or a trailing empty field? if (baos.size() > 0 // we have some data || push // we've started a field || state == ParserState.EMBEDDEDQUOTE // Just seen "" ) { list.add(baos.toString()); } } return list.toArray(new String[list.size()]); } -- You are receiving this mail because: You are the assignee for the bug.
