I counted to 15 awk calls in our latest pipeline processing drug compounds to build predictive machine learning models of them: https://github.com/pharmbio/ptp-

I counted to 15 awk calls in our latest pipeline processing drug compounds to build predictive machine learning models of them: https://github.com/pharmbio/ptp-project/blob/master/exp/20180426-wo-drugbank/wo_drugbank_wf.go https://github.com/pharmbio/ptp-project/blob/master/exp/2018... One of the most eye-opening aspects of awk (goes for other pipeable commandline tools too), was how they support iterative development of pipelines regardless of data size. We tried SQLite at some point for some of the stages because of pretty complicated selections sometimes, but it often won't give back a single result in minutes. Switching to AWK, I could immediately get some output, so I could quickly validate and iterate on the awk code until I got what I expected. The actual execution was likeways always very fast.