Closed Loyale closed 6 years ago
Very cool, thank you! And I’m happy to hear you like destiny!
Would you mind formatting the new code like the package code? Then I’ll have no spurious commits only for formatting. The indentation is tabs, and i use whitespace around <-
, =
, and after ,
.
Very elegant use of rows
and cols
😄
Should be formatted as requested now. Let me know if I can tweak anything else to match. Cheers.
Thank you, it’s perfect!
upper.tri() forcibly coerces input to a base matrix() using as.matrix(). This can be memory limiting for large sparse matrices. as.matrix() is not needed in upper.tri() to accomplish the goals of creating the mask in destiny. As such this patch defines upper.tri.sparse() in R/utils.r with exactly the same implementation as upper.tri() without the call to as.matrix(). This is replaced in R/diffusionmap.r and allows destiny to proceed on an otherwise large 10x genomics single cell dataset (>150K cells) on a 32 Gb machine. Thanks for destiny. Great package!