Dear Statalist

I have a yearly dataset of chinses non-financial listed firms from (2010-2017), I want to calculate the three-year moving sum of the absolute value of annual discretionary accruals.
I don't know the command that give me the correct result.
any type of help is considered highly appreciated.
omer Sahil
that's how my dataset look like.
id stock code years DisACC absloutDisAcc
1 2 2010
1 2 2011 0.0333704 0.0333704
1 2 2012 0.0297041 0.0297041
1 2 2013 0.0381274 0.0381274
1 2 2014 -0.0375272 0.0375272
1 2 2015 0.0132477 0.0132477
1 2 2016 -0.0121819 0.0121819
1 2 2017 -0.0285276 0.0285276
2 4 2010
2 4 2011 0.3887737 0.3887737
2 4 2012 0.313039 0.313039
2 4 2013 0.5399743 0.5399743
2 4 2014 0.2769081 0.2769081
2 4 2015 0.1409279 0.1409279
2 4 2016 -0.2453816 0.2453816
2 4 2017 0.4473724 0.4473724
3 5 2010
3 5 2011 -0.0254381 0.0254381
3 5 2012 -0.0837116 0.0837116
3 5 2013 0.0610106 0.0610106
3 5 2014 0.0911542 0.0911542
3 5 2015 -0.0009404 0.0009404
3 5 2016 0.1230942 0.1230942
3 5 2017 0.1375438 0.1375438