split data row

[belaruska]

I have a file with variable length rows

i.e.:

val1:data1;val2:data2;val3:data3
val4:data4
val5:data5;val6:data6

i need to split values and data, such that all values are in one string and all data in the other

Please advise on the ways to do it

[ray.wurlod]

Pre-process the file (with tr perhaps) to convert all the ";" characters to line terminators - that way you will have one value:data pair per line. If you use tr as the filter command on a Sequential File stage, the stage will read stdout from the tr command as the source.