This one (without use of cursor or loops opr UDF) will perform and scale much better...

declare @pInput varchar(max)

declare @n int

set @pInput = '1:1,3,5,7,4:56,43,58,5:34,67r,234'

set @pInput = ',' + @pInput + ','

select @n =LEN(@pInput)

set rowcount @n

select IDENTITY( int,1,1) as id into #tally

from sys.columns

set rowcount 0

declare @val1 varchar(500)

declare @res table (id int, vals varchar(1000), val1 varchar(500), val2 varchar(500))

insert into @res (id, vals)

select id

,SUBSTRING(@pInput,id+1,CHARINDEX(',',@pInput,id+1)-id-1) vals

from #tally

where SUBSTRING(@pInput,id,1) = ',' and id < LEN(@pInput)

update @res

set val1 = case when CHARINDEX(':', vals) > 0 then SUBSTRING(vals,0,CHARINDEX(':', vals)) else @val1 end

,val2 = case when CHARINDEX(':', vals) > 0 then SUBSTRING(vals,CHARINDEX(':', vals)+1,1000) else vals end

,@val1 = case when CHARINDEX(':', vals) > 0 then SUBSTRING(vals,0,CHARINDEX(':', vals)) else @val1 end

select val1, val2 from @res

drop table #tally

If your string to split expected to be very large, you can add clustered unique index on id column of #tally table and use # table instead of table variable for the result table. Otherwise, I think this code is fine as it is...