tMap and tJoin component wont accept 2nd connection

[Anonymous]

I have a process where I want to join the upper and lower flows using a tMap but once I connect 1 component's input to the tMap I can't connect the other. If I create a basic job with 2 ExcelFileInputs I can join 1 to a tMap and get the lookup connection for the other.
Any troubleshooting or debugging advice as to why tMap or tJoin won't accept a 2nd connection for this particular job?

[cterenzi]

Talend won't let you create a circular path in a job. Instead of replicating a flow and joining it back, use two input components with the same source.

[Anonymous]

Hiya, Thanks for helping out. I'm not sure I follow as the 2 input components are using the same file as input I just need to process row 1 differently to extract some extra data and join it with rows 3 onwards (which I validate against a schema beforehand)
If I can't join the flows together whats the best way to store temporary values I extract from the header row so I can include it in tMap

[TRF]

Hi,
You cannot separate a flow (using tReplicate or tMap for example) and reunify it (using tUnite or tMap for example) in the same subjob.
Regards,
TRF

[Anonymous]

So is there a recommended way to process a file which has the structure
Legend  

|1| = row number
Comma used to indicate separate cells in excel

|1|This report was generated at 23:20 September 2014
|2|
|3|Machine,Status,Value
|4|A, Good, 14
|5|A, Bad, 10

and end up with 

|1|Machine,Status,Value,Time,Month,Year
|2|A, Good, 14, 23:20,September,2014
|3|A, Bad, 10,23:20,September,2014

[TRF]

1- read 1st line using tFileInputMSDelimited and to retrieve hour and date as 3 separate fields of the same record stored in tHashOutput_1.
2- read file from 4th line to the end and include content of tHashOutput_1 for each record using tMap
                          tHashInput__1
                                     |
tFileInputDelimited  --> tMap --> tFileOutputDelimited
You may prefer to store result of step 1 into 3 global variables (time, month, year) then use them to populate the records in tMap (no need of tHash neither join in tMap). 
In this case, just use a tJavaRow after tFileInputMSDelimited and "globalMap.put" to store values in the variables.
Regards,
TRF

[Anonymous]

Could I put the File Inputs in serial (not parallel as they are above) and use setGlobalvars / globalMap.put before the 2nd Input?
tFileExcelInput -->ExtractVarsFromRow1 -->SetGlobalVar -->tFileExcelInput(same input settings) -->tMap (with GlobalVars)--->FileOutput                                                                                                                                    
Is there a possibility of File Locks or other issues?

[TRF]

You should prefer this design:
tFileExcelInput --> tJava ( SetGlobalVar)
 |
onSubjobOk
 |
tFileExcelInput(same input settings) -->tMap (with GlobalVars)--->FileOutput
No risk of lock and better for lisibility and maintainability.
Regards,
TRF