我一直在与cumsum战斗，以适应多种条件。在这种情况下，我需要为year，site，canopy和treatemt_abbr执行一年中每一天的运行温度总和。以下是简化数据的示例：
site year doy canopy power treatment_abbr airtemp_comb_nearby gdd10
cfc 2009 1 closed 0 h2 -18.153490 0
cfc 2009 2 closed 0 h2 18.153490 8
cfc 2009 3 closed 0 h2 13.153490 3
cfc 2009 1 open 0 ac 12.490 2
cfc 2009 2 open 0 ac 16.912620 6
hwrc 2012 1 closed 0 dc 11.146437 1
hwrc 2012 2 closed 0 dc 2.005500 0
hwrc 2012 3 closed 0 dc 2.5500 0
hwrc 2012 4 closed 0 dc 22.1234 12
hwrc 2012 5 closed 0 dc 2.005500 0
我的实际数据集非常大，所以我想使用data.table功能。在我看来，以下应该工作。它会创建新列“gddsum10”，但无法执行运行总和。知道我在这里做错了吗？
dt[order(doy), gddsum10:=cumsum(gdd10), by=c("year", "doy", "site",
"canopy", "treatment_abbr")]
我正在寻找与新列"gddsum10"：
site year doy canopy power treatment_abbr airtemp_comb_nearby gdd10 gddsum10
cfc 2009 1 closed 0 h2 -18.153490 0 0
cfc 2009 2 closed 0 h2 18.153490 8 8
cfc 2009 3 closed 0 h2 13.153490 3 11
cfc 2009 1 open 0 ac 12.490 2 2
cfc 2009 2 open 0 ac 16.912620 6 8
hwrc 2012 1 closed 0 dc 11.146437 1 1
hwrc 2012 2 closed 0 dc 2.005500 0 1
hwrc 2012 3 closed 0 dc 2.5500 0 1
hwrc 2012 4 closed 0 dc 22.1234 12 13
hwrc 2012 5 closed 0 dc 2.005500 0 13