-- instead of using y = rbind(x, y), which is O(n^2) in a loop when processing lines into a data structure in R, preallocate a matrix and explicitly assign each row to x. This results in a radical performance improvement when reading large tables into R. It's possible with this optimization to read in a 70MB table for variantQCReport.R with 200K lines for 800 samples. |
||
|---|---|---|
| .. | ||
| scripts/org/broadinstitute/sting | ||
| src/org/broadinstitute/sting/utils/R/gsalib | ||
| titvFPEst.R | ||