h2o.merge will not work with frame with 1 row of <dbl> column

[exalate-issue-sync]

With h2o.merge in R, if you try to join a single row H2O frame where the join is on a column to another frame, it will fail (below) or returns wrong number of rows:

Input:

{code:r}left <- data.frame(fruit = c(-177000000, -4000000, 100000000000, 200000000000, 1000000000000),
color <- c('red', 'orange', 'yellow', 'red', 'blue'))

right <- data.frame(fruit = c(-177000000), citrus <- c(FALSE))

left_hf <- as.h2o(left)
right_hf <- as.h2o(right)

merged <- h2o.merge(left_hf, right_hf, all.x = TRUE)
merged{code}

{quote}Output:
ERROR while rich displaying an object: Error in .h2o.doSafeREST(h2oRestApiVersion = h2oRestApiVersion, urlSuffix = page, :
ERROR MESSAGE:
java.lang.AssertionError
Traceback:

FUN(X[[i]], ...)

tryCatch(withCallingHandlers({

. if (!mime %in% names(repr::mime2repr)) . stop("No repr_* for mimetype ", mime, " in repr::mime2repr"). rpr <- repr::mime2repr[[mime]](obj). if (is.null(rpr)) . return(NULL). prepare_content(is.raw(rpr), rpr). }, error = error_handler), error = outer_handler)

tryCatchList(expr, classes, parentenv, handlers)

tryCatchOne(expr, names, parentenv, handlers[[1L]])

doTryCatch(return(expr), name, parentenv, handler)

withCallingHandlers({

. if (!mime %in% names(repr::mime2repr)) . stop("No repr_* for mimetype ", mime, " in repr::mime2repr"). rpr <- repr::mime2repr[[mime]](obj). if (is.null(rpr)) . return(NULL). prepare_content(is.raw(rpr), rpr). }, error = error_handler)

repr::mime2repr[[mime]](obj)

repr_text.default(obj)

paste(capture.output(print(obj)), collapse = "\n")

10. capture.output(print(obj))

11. evalVis(expr)

12. withVisible(eval(expr, pf))

13. eval(expr, pf)

14. eval(expr, pf)

15. print(obj)

16. print.H2OFrame(obj)

17. print(head(x, n, m))

18. head(x, n, m)

19. head.H2OFrame(x, n, m)

20. nrow

21. dim

22. dim.H2OFrame

23. .eval.frame

24. .eval.driver

25. .h2o._remoteSend(.h2o._RAPIDS, h2oRestApiVersion = 99, ast = exec_str, . session_id = h2o.getConnection()@mutable$session_id, method = "POST")

26. .h2o.doSafeREST(h2oRestApiVersion = h2oRestApiVersion, urlSuffix = page, . parms = .params, method = method)

27. stop(msg){quote}

[exalate-issue-sync]

Neema Mashayekhi commented: The issue also occurs for multiple rows (as long as the values are the same):

{code:r}left <- data.frame(fruit = c(-177000000, -4000000, 100000000000, 200000000000, 1000000000000),
color <- c('red', 'orange', 'yellow', 'red', 'blue'))
right <- data.frame(fruit = c(-177000000, -177000000),
citrus <- c(FALSE))
left_hf <- as.h2o(left)
right_hf <- as.h2o(right)
merged <- h2o.merge(left_hf, right_hf, all.x = TRUE){code}

[exalate-issue-sync]

Wendy commented: One more example from Kuba:

{noformat}[A, 12][B, 13][C, 14][D, 15]{noformat}

right frame

{noformat}[Y, 10000][B, 20000][X, 30000][D, 40000]{noformat}

Doing right marge ( allLeft=False, allRight=True) using radix method
expected:

{noformat}[Y,null,10000][B,13,20000][X,null,30000][D,15,40000]{noformat}

got:

{noformat}[Y,13,10000][B,null,20000][X,null,30000][D,12,40000]{noformat}

[h2o-ops]

JIRA Issue Migration Info

Jira Issue: PUBDEV-7362
Assignee: Wendy
Reporter: Neema Mashayekhi
State: Resolved
Fix Version: 3.30.0.4
Attachments: N/A
Development PRs: Available

Linked PRs from JIRA

#4515
#4631
h2oai/sparkling-water#2149