. which parallel
R:\STATA\ado\site\p\parallel.ado
*! version 1.19.0 26jul2017
Expected behavior and actual behavior
parallel.ado:225:
gen _`parallelid'cut = ceil(_n*${PLL_CLUSTERS}/_N) //each of size _N/$PLL_CLUSTERS
can return values > $PLL_CLUSTERS,
thus when $PLL_CLUSTERS numbers of subsets are combined the combined data have less observations than the origional data.
The example below show how what is return for clusters 7, 11, 14 with N=2771050. An alternative is shown below the value returned to _`parallelid'cut
local _N 2771050
local _n = `_N'
qui foreach c of numlist 7 11 14 {
noi di _dup(72)"-"
global PLL_CLUSTERS = `c'
noi di "parallel.ado:225:" ceil(`_n'*${PLL_CLUSTERS}/`_N')
noi di "alternative :" min( `c', ceil(`_n'*${PLL_CLUSTERS}/`_N'))
}
------------------------------------------------------------------------
parallel.ado:225:8
alternative :7
------------------------------------------------------------------------
parallel.ado:225:12
alternative :11
------------------------------------------------------------------------
parallel.ado:225:15
alternative :14
. which parallel R:\STATA\ado\site\p\parallel.ado *! version 1.19.0 26jul2017
Expected behavior and actual behavior
parallel.ado:225:
gen _`parallelid'cut = ceil(_n*${PLL_CLUSTERS}/_N) //each of size _N/$PLL_CLUSTERS
can return values > $PLL_CLUSTERS,
thus when $PLL_CLUSTERS numbers of subsets are combined the combined data have less observations than the origional data.
The example below show how what is return for clusters 7, 11, 14 with N=2771050. An alternative is shown below the value returned to _`parallelid'cut
local _N 2771050
local _n = `_N'
qui foreach c of numlist 7 11 14 {
}