Stata code needed

I have a dataset of conflict disputes. They have unique identifiers. It is a directed dyadic country year dataset. But I need it to be a directed dyadic dataset. In other words, a dispute could involved three countries, such as the US, Taiwan, and China across, say, two years. The dispute began in 1949 between Taiwan and China and ended in 1950. The US got involved in 1950. So there are currently six observations. US-China 1950, China-US 1950, China-Taiwan 1949, Taiwan-China 1949, China-Taiwan 1950, Taiwan-China 1950. I need only the last year of observations for the data pairs for each dispute. So I would need US-China 1950, China-US 1950, China-Taiwan 1950, and Taiwan-China 1950. It is a straightforward problem that only needs like three lines of code. I just can't remember how I did it in Stata. I can check the data so the coder doesn't have to worry about that. The dataset is attached below, as is the codebook.