Shell Programming and Scripting

I need to split a file into n number of files but with one condition.
Similar field 1 lines should end up in a single file.

Eg:
XYZ|sdfwe|fsd|SDFDW
ABC|sdfdsf|werw|ASD
ZXC|awrwe|wrfwe|WEq
XYZ|werwe|QWEQW|SADQ
SER|asq|SADEW|AS|asa
ABC|asdwq|sdad|SAdf
BCG|dfg|sdfa|fgdsf

if n=2
file1:
XYZ|sdfwe|fsd|SDFDW
XYZ|werwe|QWEQW|SADQ
ZXC|awrwe|wrfwe|WEq

file2
ABC|sdfdsf|werw|ASD
ABC|asdwq|sdad|SAdf
SER|asq|SADEW|AS|asa
BCG|dfg|sdfa|fgdsf

couple of questions:

my problem is splitting the file exactly into "n" numbers. split command split the files based on either with no of lines or the size given. If the number of lines is odd, i am not able to split it into exactly the n number.

Selection into the split files is not an issue.Only thing is the lines with the same field 1 values (lines are delimited by '|') should end up in a single file.

In my example i showed lines start with 'XYZ' end in the same splitted file.

why is 'ZXC' in the same file with 'XYZ'?

It can be in any one of the splitted file. But the thing is if any other line starts with 'ZXC' presents in main file, both should be end in the same splitted file.

ok, so.... how do you decide what line goes to what file?
E.g. how do you know that 'XYZ' should go to file1 and 'ABC' should go to 'file2'?
Can I have both the 'XYZ' and the 'ABC' lines in the same file? I guess I don't understand the algorithm here.
And for the same: if 'n==3', how do you know what goes where?

There is no matter whether 'XYZ' and 'ABC' end up in the same or different file. Also there is no condition for the lines to go into any specific file.

Only condition is lines starting with the same field should end up in the same splitted file. And the number of splitted file is predefined i.e. 'n'.

But if suppose there are 10 lines in a file all 10 lines starts with 'XYZ' there wont be any splitting. And if suppoes 9 XYZ lins and 1 ABC lines in main file it should be split into two files with all XYZ in one file and ABC in another file. So there will be two splitted file eventhough if split count is more than 2.

so to put in another words the resultant number of split files can be 1 to max of predined 'n'.

I am thinking in moving the lines from the main file into the split files by roundrobin manner. That is first get all unique field 1 values. For each value move the corresponding lines into the split files in a roundrobin manner. But i am not sure whether i am complicating myself?

Hope this explanation is not confusing!!!