add h2o for internal tests

_control/solutions.csv

@@ -16,3 +16,5 @@ spark,join
 clickhouse,groupby
 cudf,groupby
 cudf,join
+h2o,groupby
+h2o,join

_launcher/launcher.R

@@ -14,7 +14,7 @@ readret = function(x) {
 file.ext = function(x) {
   ans = switch(
     x,
-    "data.table"=, "dplyr"="R",
+    "data.table"=, "dplyr"=, "h2o"="R",
     "pandas"=, "cudf"=, "spark"=, "pydatatable"=, "modin"=, "dask"="py",
     "clickhouse"="sql",
     "juliadf"="jl"

_launcher/solution.R

@@ -109,7 +109,7 @@ if ("quiet" %in% names(args)) {
 file.ext = function(x) {
   ans = switch(
     x,
-    "data.table"=, "dplyr"="R",
+    "data.table"=, "dplyr"=, "h2o"="R",
     "pandas"=, "cudf"=, "spark"=, "pydatatable"=, "modin"=, "dask"="py",
     "clickhouse"="sql",
     "juliadf"="jl"
@@ -169,7 +169,7 @@ setenv(data_name_env, d)
 
 ns = solution.path(s)
 ext = file.ext(s)
-localcmd = if (ext=="sql") { # sql scripts are using extra exec shell script, related only to clickhouse as of now
+localcmd = if (s %in% c("clickhouse","h2o")) { # custom launcher bash script, for clickhouse and h2o
   sprintf("exec.sh %s %s", t, d)
 } else sprintf("%s-%s.%s", t, ns, ext)
 cmd = sprintf("./%s/%s", ns, localcmd)

h2o/exec.sh

@@ -0,0 +1,28 @@
+#!/bin/bash
+set -e
+
+if [ "$#" -ne 2 ]; then
+    echo "usage: ./h2o/exec.sh groupby G1_1e7_1e2_0_0";
+    exit 1
+fi;
+
+source ./h2o/h2o.sh
+
+h2o_active || echo "h2o instance should not be already running, investigate" >&2
+h2o_active || exit 1
+
+# start h2o
+h2o_start "h2o_$1_$2"
+
+# confirm h2o working
+h2o_active || sleep 30
+h2o_active || echo "h2o instance should be already running, investigate" >&2
+h2o_active || exit 1
+
+# execute benchmark script
+./h2o/$1-h2o.R || echo "# h2o/exec.sh: benchmark script for $2 terminated with error" >&2
+
+# stop h2o instance
+h2o_stop && echo "# h2o/exec.sh: stopping h2o instance finished" || echo "# h2o/exec.sh: stopping h2o instance failed" >&2
+h2o_active || exit 1
+

h2o/groupby-h2o.R

@@ -0,0 +1,142 @@
+#!/usr/bin/env Rscript
+
+cat("# groupby-h2o.R\n")
+
+source("./_helpers/helpers.R")
+
+suppressPackageStartupMessages(library("h2o", lib.loc="./h2o/r-h2o", warn.conflicts=FALSE, quietly=TRUE))
+ver = packageVersion("h2o")
+git = ""
+task = "groupby"
+solution = "h2o"
+fun = "h2o.group_by"
+cache = TRUE
+on_disk = FALSE
+
+h = h2o.init(startH2O=FALSE, port=55888)
+h2o.no_progress()
+
+data_name = Sys.getenv("SRC_GRP_LOCAL")
+src_grp = file.path("data", paste(data_name, "csv", sep="."))
+cat(sprintf("loading dataset %s\n", data_name))
+
+x = h2o.importFile(src_grp, col.types=c("string","string","string","int","int","int","int","int","real"))
+# using string due to ERROR caused by water.parser.ParseDataset$H2OParseException: Exceeded categorical limit on column #3 (using 1-based indexing).  Consider reparsing this column as a string.
+# https://0xdata.atlassian.net/browse/PUBDEV-7533
+print(nrow(x))
+
+task_init = proc.time()[["elapsed"]]
+cat("grouping...\n")
+
+question = "sum v1 by id1" # q1
+t = system.time(print(dim(ans<-h2o.group_by(x, by="id1", sum("v1")))))[["elapsed"]]
+m = memory_usage()
+chkt = system.time(chk<-list(v1=sum(ans[["sum_v1"]])))[["elapsed"]]
+write.log(run=1L, task=task, data=data_name, in_rows=nrow(x), question=question, out_rows=nrow(ans), out_cols=ncol(ans), solution=solution, version=ver, git=git, fun=fun, time_sec=t, mem_gb=m, cache=cache, chk=make_chk(chk), chk_time_sec=chkt, on_disk=on_disk)
+h2o.rm(ans)
+t = system.time(print(dim(ans<-h2o.group_by(x, by="id1", sum("v1")))))[["elapsed"]]
+m = memory_usage()
+chkt = system.time(chk<-list(v1=sum(ans[["sum_v1"]])))[["elapsed"]]
+write.log(run=2L, task=task, data=data_name, in_rows=nrow(x), question=question, out_rows=nrow(ans), out_cols=ncol(ans), solution=solution, version=ver, git=git, fun=fun, time_sec=t, mem_gb=m, cache=cache, chk=make_chk(chk), chk_time_sec=chkt, on_disk=on_disk)
+print(head(ans, 3))
+print(tail(ans, 3))
+h2o.rm(ans)
+
+question = "sum v1 by id1:id2" # q2
+t = system.time(print(dim(ans<-h2o.group_by(x, by=c("id1","id2"), sum("v1")))))[["elapsed"]]
+m = memory_usage()
+chkt = system.time(chk<-list(v1=sum(ans[["sum_v1"]])))[["elapsed"]]
+write.log(run=1L, task=task, data=data_name, in_rows=nrow(x), question=question, out_rows=nrow(ans), out_cols=ncol(ans), solution=solution, version=ver, git=git, fun=fun, time_sec=t, mem_gb=m, cache=cache, chk=make_chk(chk), chk_time_sec=chkt, on_disk=on_disk)
+h2o.rm(ans)
+t = system.time(print(dim(ans<-h2o.group_by(x, by=c("id1","id2"), sum("v1")))))[["elapsed"]]
+m = memory_usage()
+chkt = system.time(chk<-list(v1=sum(ans[["sum_v1"]])))[["elapsed"]]
+write.log(run=2L, task=task, data=data_name, in_rows=nrow(x), question=question, out_rows=nrow(ans), out_cols=ncol(ans), solution=solution, version=ver, git=git, fun=fun, time_sec=t, mem_gb=m, cache=cache, chk=make_chk(chk), chk_time_sec=chkt, on_disk=on_disk)
+print(head(ans, 3))
+print(tail(ans, 3))
+h2o.rm(ans)
+
+question = "sum v1 mean v3 by id3" # q3
+t = system.time(print(dim(ans<-h2o.group_by(x, by="id3", sum("v1"), mean("v3")))))[["elapsed"]]
+m = memory_usage()
+chkt = system.time(chk<-list(v1=sum(ans[["sum_v1"]]), v3=sum(ans[["mean_v3"]])))[["elapsed"]]
+write.log(run=1L, task=task, data=data_name, in_rows=nrow(x), question=question, out_rows=nrow(ans), out_cols=ncol(ans), solution=solution, version=ver, git=git, fun=fun, time_sec=t, mem_gb=m, cache=cache, chk=make_chk(chk), chk_time_sec=chkt, on_disk=on_disk)
+h2o.rm(ans)
+t = system.time(print(dim(ans<-h2o.group_by(x, by="id3", sum("v1"), mean("v3")))))[["elapsed"]]
+m = memory_usage()
+chkt = system.time(chk<-list(v1=sum(ans[["sum_v1"]]), v3=sum(ans[["mean_v3"]])))[["elapsed"]]
+write.log(run=2L, task=task, data=data_name, in_rows=nrow(x), question=question, out_rows=nrow(ans), out_cols=ncol(ans), solution=solution, version=ver, git=git, fun=fun, time_sec=t, mem_gb=m, cache=cache, chk=make_chk(chk), chk_time_sec=chkt, on_disk=on_disk)
+print(head(ans, 3))
+print(tail(ans, 3))
+h2o.rm(ans)
+
+question = "mean v1:v3 by id4" # q4
+t = system.time(print(dim(ans<-h2o.group_by(x, by="id4", mean("v1"), mean("v2"), mean("v3")))))[["elapsed"]]
+m = memory_usage()
+chkt = system.time(chk<-list(v1=sum(ans[["mean_v1"]]), v2=sum(ans[["mean_v2"]]), v3=sum(ans[["mean_v3"]])))[["elapsed"]]
+write.log(run=1L, task=task, data=data_name, in_rows=nrow(x), question=question, out_rows=nrow(ans), out_cols=ncol(ans), solution=solution, version=ver, git=git, fun=fun, time_sec=t, mem_gb=m, cache=cache, chk=make_chk(chk), chk_time_sec=chkt, on_disk=on_disk)
+h2o.rm(ans)
+t = system.time(print(dim(ans<-h2o.group_by(x, by="id4", mean("v1"), mean("v2"), mean("v3")))))[["elapsed"]]
+m = memory_usage()
+chkt = system.time(chk<-list(v1=sum(ans[["mean_v1"]]), v2=sum(ans[["mean_v2"]]), v3=sum(ans[["mean_v3"]])))[["elapsed"]]
+write.log(run=2L, task=task, data=data_name, in_rows=nrow(x), question=question, out_rows=nrow(ans), out_cols=ncol(ans), solution=solution, version=ver, git=git, fun=fun, time_sec=t, mem_gb=m, cache=cache, chk=make_chk(chk), chk_time_sec=chkt, on_disk=on_disk)
+print(head(ans, 3))
+print(tail(ans, 3))
+h2o.rm(ans)
+
+question = "sum v1:v3 by id6" # q5
+t = system.time(print(dim(ans<-h2o.group_by(x, by="id6", sum("v1"), sum("v2"), sum("v3")))))[["elapsed"]]
+m = memory_usage()
+chkt = system.time(chk<-list(v1=sum(ans[["sum_v1"]]), v2=sum(ans[["sum_v2"]]), v3=sum(ans[["sum_v3"]])))[["elapsed"]]
+write.log(run=1L, task=task, data=data_name, in_rows=nrow(x), question=question, out_rows=nrow(ans), out_cols=ncol(ans), solution=solution, version=ver, git=git, fun=fun, time_sec=t, mem_gb=m, cache=cache, chk=make_chk(chk), chk_time_sec=chkt, on_disk=on_disk)
+h2o.rm(ans)
+t = system.time(print(dim(ans<-h2o.group_by(x, by="id6", sum("v1"), sum("v2"), sum("v3")))))[["elapsed"]]
+m = memory_usage()
+chkt = system.time(chk<-list(v1=sum(ans[["sum_v1"]]), v2=sum(ans[["sum_v2"]]), v3=sum(ans[["sum_v3"]])))[["elapsed"]]
+write.log(run=2L, task=task, data=data_name, in_rows=nrow(x), question=question, out_rows=nrow(ans), out_cols=ncol(ans), solution=solution, version=ver, git=git, fun=fun, time_sec=t, mem_gb=m, cache=cache, chk=make_chk(chk), chk_time_sec=chkt, on_disk=on_disk)
+print(head(ans, 3))
+print(tail(ans, 3))
+h2o.rm(ans)
+
+question = "median v3 sd v3 by id4 id5" # q6
+t = system.time(print(dim(ans<-h2o.group_by(x, by=c("id4","id5"), median("v3"), sd("v3")))))[["elapsed"]]
+m = memory_usage()
+chkt = system.time(chk<-list(sum(ans[["median_v3"]]), sum(ans[["sd_v3"]])))[["elapsed"]]
+write.log(run=1L, task=task, data=data_name, in_rows=nrow(x), question=question, out_rows=nrow(ans), out_cols=ncol(ans), solution=solution, version=ver, git=git, fun=fun, time_sec=t, mem_gb=m, cache=cache, chk=make_chk(chk), chk_time_sec=chkt, on_disk=on_disk)
+h2o.rm(ans)
+t = system.time(print(dim(ans<-h2o.group_by(x, by=c("id4","id5"), median("v3"), sd("v3")))))[["elapsed"]]
+m = memory_usage()
+chkt = system.time(chk<-list(sum(ans[["median_v3"]]), sum(ans[["sd_v3"]])))[["elapsed"]]
+write.log(run=2L, task=task, data=data_name, in_rows=nrow(x), question=question, out_rows=nrow(ans), out_cols=ncol(ans), solution=solution, version=ver, git=git, fun=fun, time_sec=t, mem_gb=m, cache=cache, chk=make_chk(chk), chk_time_sec=chkt, on_disk=on_disk)
+print(head(ans, 3))
+print(tail(ans, 3))
+h2o.rm(ans)
+
+question = "max v1 - min v2 by id3" # q7
+
+
+question = "largest two v3 by id6" # q8
+
+
+question = "regression v1 v2 by id2 id4" # q9
+
+
+question = "sum v3 count by id1:id6" # q10
+t = system.time(print(dim(ans<-h2o.group_by(x, by=c("id1","id2","id3","id4","id5","id6"), sum("v3"), nrow("id1","id2","id3","id4","id5","id6")))))[["elapsed"]]
+m = memory_usage()
+chkt = system.time(chk<-list(sum(ans[["sum_v3"]]), sum(ans[["nrow"]])))[["elapsed"]]
+write.log(run=1L, task=task, data=data_name, in_rows=nrow(x), question=question, out_rows=nrow(ans), out_cols=ncol(ans), solution=solution, version=ver, git=git, fun=fun, time_sec=t, mem_gb=m, cache=cache, chk=make_chk(chk), chk_time_sec=chkt, on_disk=on_disk)
+h2o.rm(ans)
+t = system.time(print(dim(ans<-h2o.group_by(x, by=c("id1","id2","id3","id4","id5","id6"), sum("v3"), nrow("id1","id2","id3","id4","id5","id6")))))[["elapsed"]]
+m = memory_usage()
+chkt = system.time(chk<-list(sum(ans[["sum_v3"]]), sum(ans[["nrow"]])))[["elapsed"]]
+write.log(run=2L, task=task, data=data_name, in_rows=nrow(x), question=question, out_rows=nrow(ans), out_cols=ncol(ans), solution=solution, version=ver, git=git, fun=fun, time_sec=t, mem_gb=m, cache=cache, chk=make_chk(chk), chk_time_sec=chkt, on_disk=on_disk)
+print(head(ans, 3))
+print(tail(ans, 3))
+h2o.rm(ans)
+
+h2o.removeAll()
+
+cat(sprintf("grouping finished, took %.0fs\n", proc.time()[["elapsed"]]-task_init))
+
+if (!interactive()) q("no", status=0)

h2o/h2o.sh

@@ -0,0 +1,20 @@
+h2o_start() {
+  ((!$#)) && echo "h2o_start require h2o instance name as a parameter" >&2 && return 1
+  echo '# h2o_start: starting h2o instance'
+  nohup java -Xmx100G -Xms100G -cp ./h2o/r-h2o/h2o/java/h2o.jar water.H2OApp -name "$1" -baseport 55888 > ./h2o/log/$1.out 2> ./h2o/log/$1.err < /dev/null &
+  sleep 10
+}
+h2o_stop() {
+  echo '# h2o_stop: stopping h2o instance'
+  pidof java > /dev/null 2>&1 && killall -2 java > /dev/null 2>&1
+  sleep 2 && pidof java > /dev/null 2>&1 && sleep 15
+  pidof java > /dev/null 2>&1 && killall -15 java > /dev/null 2>&1
+  sleep 2 && pidof java > /dev/null 2>&1 && sleep 30
+  pidof java > /dev/null 2>&1 && killall -9 java > /dev/null 2>&1
+  sleep 2 && pidof java > /dev/null 2>&1 && sleep 60 && pidof java > /dev/null 2>&1 && echo "h2o instance could not be stopped" >&2 && return 1
+  return 0
+}
+h2o_active() {
+  pidof java > /dev/null 2>&1 && curl -X GET "localhost:55888/3/About" -H "accept: application/json" > /dev/null 2>&1
+}
+

h2o/init-h2o.sh

@@ -0,0 +1,6 @@
+#!/bin/bash
+set -e
+
+# upgrade to latest stable from h2o repo
+echo 'upgrading h2o...'
+Rscript -e 'ap=available.packages(repos="http://h2o-release.s3.amazonaws.com/h2o/latest_stable_R", method="curl"); if (ap["h2o","Version"]!=packageVersion("h2o", lib.loc="./h2o/r-h2o")) update.packages(lib.loc="./h2o/r-h2o", repos="http://h2o-release.s3.amazonaws.com/h2o/latest_stable_R", method="curl", ask=FALSE, checkBuilt=TRUE, quiet=TRUE)'

h2o/join-h2o.R

@@ -0,0 +1,111 @@
+#!/usr/bin/env Rscript
+
+cat("# join-h2o.R\n")
+
+source("./_helpers/helpers.R")
+
+suppressPackageStartupMessages(library("h2o", lib.loc="./h2o/r-h2o", warn.conflicts=FALSE, quietly=TRUE))
+ver = packageVersion("h2o")
+git = ""
+task = "join"
+solution = "h2o"
+fun = "h2o.merge"
+cache = TRUE
+on_disk = FALSE
+
+h = h2o.init(startH2O=FALSE, port=55888)
+h2o.no_progress()
+
+data_name = Sys.getenv("SRC_JN_LOCAL")
+src_jn_x = file.path("data", paste(data_name, "csv", sep="."))
+y_data_name = join_to_tbls(data_name)
+src_jn_y = setNames(file.path("data", paste(y_data_name, "csv", sep=".")), names(y_data_name))
+stopifnot(length(src_jn_y)==3L)
+cat(sprintf("loading datasets %s\n", paste(c(data_name, y_data_name), collapse=", ")))
+
+x = h2o.importFile(src_jn_x, col.types=c("int","int","int","enum","enum","string","real"))
+print(nrow(x))
+small = h2o.importFile(src_jn_y[1L], col.types=c("int","enum","real"))
+medium = h2o.importFile(src_jn_y[2L], col.types=c("int","int","enum","enum","real"))
+big = h2o.importFile(src_jn_y[3L], col.types=c("int","int","int","enum","enum","string","real"))
+sapply(sapply(list(small, medium, big), nrow), print) -> nul
+
+task_init = proc.time()[["elapsed"]]
+cat("joining...\n")
+
+question = "small inner on int" # q1
+
+t = system.time(print(dim(ans<-h2o.merge(x, small, by="id1"))))[["elapsed"]]
+m = memory_usage()
+chkt = system.time(chk<-list(sum(ans[["v1"]]), sum(ans[["v2"]])))[["elapsed"]]
+write.log(run=1L, task=task, data=data_name, in_rows=nrow(x), question=question, out_rows=nrow(ans), out_cols=ncol(ans), solution=solution, version=ver, git=git, fun=fun, time_sec=t, mem_gb=m, cache=cache, chk=make_chk(chk), chk_time_sec=chkt, on_disk=on_disk)
+h2o.rm(ans)
+t = system.time(print(dim(ans<-h2o.merge(x, small, by="id1"))))[["elapsed"]]
+m = memory_usage()
+chkt = system.time(chk<-list(sum(ans[["v1"]]), sum(ans[["v2"]])))[["elapsed"]]
+write.log(run=2L, task=task, data=data_name, in_rows=nrow(x), question=question, out_rows=nrow(ans), out_cols=ncol(ans), solution=solution, version=ver, git=git, fun=fun, time_sec=t, mem_gb=m, cache=cache, chk=make_chk(chk), chk_time_sec=chkt, on_disk=on_disk)
+print(head(ans, 3))
+print(tail(ans, 3))
+h2o.rm(ans)
+
+question = "medium inner on int" # q2
+t = system.time(print(dim(ans<-h2o.merge(x, medium, by="id2"))))[["elapsed"]]
+m = memory_usage()
+chkt = system.time(chk<-list(sum(ans[["v1"]]), sum(ans[["v2"]])))[["elapsed"]]
+write.log(run=1L, task=task, data=data_name, in_rows=nrow(x), question=question, out_rows=nrow(ans), out_cols=ncol(ans), solution=solution, version=ver, git=git, fun=fun, time_sec=t, mem_gb=m, cache=cache, chk=make_chk(chk), chk_time_sec=chkt, on_disk=on_disk)
+h2o.rm(ans)
+t = system.time(print(dim(ans<-h2o.merge(x, medium, by="id2"))))[["elapsed"]]
+m = memory_usage()
+chkt = system.time(chk<-list(sum(ans[["v1"]]), sum(ans[["v2"]])))[["elapsed"]]
+write.log(run=2L, task=task, data=data_name, in_rows=nrow(x), question=question, out_rows=nrow(ans), out_cols=ncol(ans), solution=solution, version=ver, git=git, fun=fun, time_sec=t, mem_gb=m, cache=cache, chk=make_chk(chk), chk_time_sec=chkt, on_disk=on_disk)
+print(head(ans, 3))
+print(tail(ans, 3))
+h2o.rm(ans)
+
+question = "medium outer on int" # q3
+t = system.time(print(dim(ans<-h2o.merge(x, medium, by="id2", all.x=TRUE))))[["elapsed"]]
+m = memory_usage()
+chkt = system.time(chk<-list(sum(ans[["v1"]]), sum(ans[["v2"]], na.rm=TRUE)))[["elapsed"]]
+write.log(run=1L, task=task, data=data_name, in_rows=nrow(x), question=question, out_rows=nrow(ans), out_cols=ncol(ans), solution=solution, version=ver, git=git, fun=fun, time_sec=t, mem_gb=m, cache=cache, chk=make_chk(chk), chk_time_sec=chkt, on_disk=on_disk)
+h2o.rm(ans)
+t = system.time(print(dim(ans<-h2o.merge(x, medium, by="id2", all.x=TRUE))))[["elapsed"]]
+m = memory_usage()
+chkt = system.time(chk<-list(sum(ans[["v1"]]), sum(ans[["v2"]], na.rm=TRUE)))[["elapsed"]]
+write.log(run=2L, task=task, data=data_name, in_rows=nrow(x), question=question, out_rows=nrow(ans), out_cols=ncol(ans), solution=solution, version=ver, git=git, fun=fun, time_sec=t, mem_gb=m, cache=cache, chk=make_chk(chk), chk_time_sec=chkt, on_disk=on_disk)
+print(head(ans, 3))
+print(tail(ans, 3))
+h2o.rm(ans)
+
+question = "medium inner on factor" # q4
+t = system.time(print(dim(ans<-h2o.merge(x, medium, by="id5"))))[["elapsed"]]
+m = memory_usage()
+chkt = system.time(chk<-list(sum(ans[["v1"]]), sum(ans[["v2"]])))[["elapsed"]]
+write.log(run=1L, task=task, data=data_name, in_rows=nrow(x), question=question, out_rows=nrow(ans), out_cols=ncol(ans), solution=solution, version=ver, git=git, fun=fun, time_sec=t, mem_gb=m, cache=cache, chk=make_chk(chk), chk_time_sec=chkt, on_disk=on_disk)
+h2o.rm(ans)
+t = system.time(print(dim(ans<-h2o.merge(x, medium, by="id5"))))[["elapsed"]]
+m = memory_usage()
+chkt = system.time(chk<-list(sum(ans[["v1"]]), sum(ans[["v2"]])))[["elapsed"]]
+write.log(run=2L, task=task, data=data_name, in_rows=nrow(x), question=question, out_rows=nrow(ans), out_cols=ncol(ans), solution=solution, version=ver, git=git, fun=fun, time_sec=t, mem_gb=m, cache=cache, chk=make_chk(chk), chk_time_sec=chkt, on_disk=on_disk)
+print(head(ans, 3))
+print(tail(ans, 3))
+h2o.rm(ans)
+
+question = "big inner on int" # q5
+t = system.time(print(dim(ans<-h2o.merge(x, big, by="id3"))))[["elapsed"]]
+m = memory_usage()
+chkt = system.time(chk<-list(sum(ans[["v1"]]), sum(ans[["v2"]])))[["elapsed"]]
+write.log(run=1L, task=task, data=data_name, in_rows=nrow(x), question=question, out_rows=nrow(ans), out_cols=ncol(ans), solution=solution, version=ver, git=git, fun=fun, time_sec=t, mem_gb=m, cache=cache, chk=make_chk(chk), chk_time_sec=chkt, on_disk=on_disk)
+h2o.rm(ans)
+t = system.time(print(dim(ans<-h2o.merge(x, big, by="id3"))))[["elapsed"]]
+m = memory_usage()
+chkt = system.time(chk<-list(sum(ans[["v1"]]), sum(ans[["v2"]])))[["elapsed"]]
+write.log(run=2L, task=task, data=data_name, in_rows=nrow(x), question=question, out_rows=nrow(ans), out_cols=ncol(ans), solution=solution, version=ver, git=git, fun=fun, time_sec=t, mem_gb=m, cache=cache, chk=make_chk(chk), chk_time_sec=chkt, on_disk=on_disk)
+print(head(ans, 3))
+print(tail(ans, 3))
+h2o.rm(ans)
+
+h2o.removeAll()
+
+cat(sprintf("joining finished, took %.0fs\n", proc.time()[["elapsed"]]-task_init))
+
+if (!interactive()) q("no", status=0)

h2o/setup-h2o.sh

@@ -0,0 +1,5 @@
+mkdir -p ./h2o/log
+# install h2o
+mkdir -p ./h2o/r-h2o
+Rscript -e 'install.packages(c("RCurl","jsonlite"), repos="https://cloud.r-project.org", lib="./h2o/r-h2o"); install.packages("h2o", repos="http://h2o-release.s3.amazonaws.com/h2o/latest_stable_R", method="curl", lib="./h2o/r-h2o")'
+

h2o/ver-h2o.sh

@@ -0,0 +1,4 @@
+#!/bin/bash
+set -e
+
+Rscript -e 'v=read.dcf(system.file(package="h2o", lib.loc="./h2o/r-h2o", "DESCRIPTION"), fields=c("Version","Revision")); cnafill=function(x) {x=c(x); x[is.na(x)]=""; x}; fw=function(f, v) writeLines(v, file.path("h2o", f)); invisible(mapply(fw, toupper(colnames(v)), cnafill(v)))'