Getting certain variable names from a sparse matrix-CodePudding

I have a sparse matrix as follows (resulting from a glmnet::lasso)

sp_mat <- new("dgCMatrix", i = c(0L, 1L, 2L, 3L, 4L, 5L, 7L, 9L, 10L, 11L, 
13L, 14L, 16L, 17L, 18L, 19L, 20L, 21L, 22L, 24L, 25L, 26L, 28L, 
30L, 31L, 32L, 33L, 34L, 35L, 36L, 37L, 39L, 40L, 41L, 42L, 43L, 
44L, 45L, 46L, 47L, 49L, 50L, 51L, 52L), p = c(0L, 44L), Dim = c(54L, 
1L), Dimnames = list(c("Avar", "Bvar", "Cvar", "m14var", "Dvar", "a26var", "a27var", 
"k05var", "b02var", "b03var", "b04var", "b05var", "b06var", "b06var", 
"Evar", "Fvar", "Gvar", "a41var", "b01var", "Hvar", "d01var", 
"d02var", "d03var", "d04var", "d05var", "d06var", "d07var", "d08var", 
"d09var", "d10var", "d11var", "d13var", "c04var", "c05var", "c06var", 
"c07var", "c08var", "c09var", "c10var", "c11var", "c12var", "c13var", 
"c14var", "c16var", "c17var", "c20var", "c21var", "c22var", "c23var", 
"c25var", "h12var", "Ivar", "Jvar", "Kvar"), "s0"), x = c(676.349860381008, 0.0000415982516505796, -0.340580369360737, 
2.46714903242688, 0.160264243823403, 0.0833642969185734, 4.0007995406493, 
-0.0969035782630975, 0.178346528406363, 1.14886155628414, 0.130340223213916, 
3.20266511712085, -3.71593684240651, 0.00000519996597479316, 
-0.0192641833629383, -0.00170895248265324, -11.0934095084226, 
0.320500528908295, -5.20396508589174, -12.7356151414996, -3.83078700482274, 
41181.4262207653, -0.0270280830604277, -0.0492721582439391, -0.00630260482614332, 
-0.38526832789847, -0.113757419772693, -0.035812506094549, -0.0464701900255123, 
-0.154718592743792, 0.0305041274157989, -0.496893176155956, -0.0510692528769553, 
-0.254008558243732, 0.12207111760977, -0.22426734553163, 0.110177476530126, 
49.4077380004606, 0.695834880415197, 0.110469954763839, -0.113121348884623, 
0.949731767706756, 0.469379688615487, -1.25529553988954), factors = list())

54 x 1 sparse Matrix of class "dgCMatrix"
                       s0
Avar     676.349860381008
Bvar       0.000041598252
Cvar      -0.340580369361
m14var     2.467149032427
Dvar       0.160264243823
a26var     0.083364296919
a27var     .             
k05var     4.000799540649
b02var     .             
b03var    -0.096903578263
...                   ...
...                   ...
...                   ...
Jvar      -1.255295539890
Kvar       .

I would simply like to extract all variable names which are not .. So Avar until a26var, but not a27var. I am not very familiar with sparse matrices and everything I tried gave me unexpected results (such as using sp_mat[[1]], names(sp_mat) etc.).

CodePudding user response：

Would summing by rows and eliminating those that are zero work?

> names(sp_mat[rowSums(sp_mat) != 0, ])
 [1] "Avar"   "Bvar"   "Cvar"   "m14var" "Dvar"   "a26var" "k05var" "b03var"
 [9] "b04var" "b05var" "b06var" "Evar"   "Gvar"   "a41var" "b01var" "Hvar"  
[17] "d01var" "d02var" "d03var" "d05var" "d06var" "d07var" "d09var" "d11var"
[25] "d13var" "c04var" "c05var" "c06var" "c07var" "c08var" "c09var" "c11var"
[33] "c12var" "c13var" "c14var" "c16var" "c17var" "c20var" "c21var" "c22var"
[41] "c25var" "h12var" "Ivar"   "Jvar"