[TEST] Run all Travis tests with Arrow enabled

[javierluraschi]

Temp PR to investigate all Travis test to run using Arrow.

══ testthat results  ══════════════════════════════════════════════════════════════════════════════════════════════════
OK: 1848 SKIPPED: 28 FAILED: 17
1. Failure: 'sdf_bind_rows' handles column type upcasting (#804) (@test-binds.R#135) 
2. Failure: 'sdf_bind_rows' handles column type upcasting (#804) (@test-binds.R#139) 
3. Failure: gaussian_mixture.tidy() works (@test-broom-gaussian_mixture.R#14) 
4. Failure: kmeans.tidy() works (@test-broom-kmeans.R#14) 
5. Failure: random_forest.tidy() works (@test-broom-random_forest.R#15) 
6. Failure: random_forest.tidy() works (@test-broom-random_forest.R#24) 
7. Error: we can separate struct columns (#690) (@test-column-extraction.R#92) 
8. Failure: top_n works as expected (@test-dplyr-top-n.R#23) 
9. Error: ml_recommend() works (@test-ml-als.R#44) 
1. ...

Some of the failures...

── 1. Failure: 'sdf_bind_rows' handles column type upcasting (#804) (@test-binds.R#135)  ──────────────────────────────
bind_rows(df5a, df6a) not equal to sdf_bind_rows(df5a_tbl, df6a_tbl) %>% collect().
Rows in x but not y: 2. Rows in y but not x: 2. 

── 2. Failure: 'sdf_bind_rows' handles column type upcasting (#804) (@test-binds.R#139)  ──────────────────────────────
bind_rows(df6a, df5a) not equal to sdf_bind_rows(df6a_tbl, df5a_tbl) %>% collect().
Rows in x but not y: 4. Rows in y but not x: 4. 

── 3. Failure: gaussian_mixture.tidy() works (@test-broom-gaussian_mixture.R#14)  ─────────────────────────────────────
td1$size not equal to c(3, 14, 4, 11).
4/4 mismatches (average diff: 3)
[1]  0 -  3 == -3
[2] 11 - 14 == -3
[3]  7 -  4 ==  3
[4] 14 - 11 ==  3

── 4. Failure: kmeans.tidy() works (@test-broom-kmeans.R#14)  ─────────────────────────────────────────────────────────
td1$size not equal to c(14, 2, 6, 10).
4/4 mismatches (average diff: 4.5)
[1] 10 - 14 == -4
[2]  3 -  2 ==  1
[3] 14 -  6 ==  8
[4]  5 - 10 == -5

── 5. Failure: random_forest.tidy() works (@test-broom-random_forest.R#15)  ───────────────────────────────────────────
td1$importance not equal to c(0.941, 0.0586).
2/2 mismatches (average diff: 0.00487)
[1] 0.9363 - 0.9410 == -0.00467
[2] 0.0637 - 0.0586 ==  0.00507

── 6. Failure: random_forest.tidy() works (@test-broom-random_forest.R#24)  ───────────────────────────────────────────
td2$importance not equal to c(0.658, 0.342).
2/2 mismatches (average diff: 0.0163)
[1] 0.642 - 0.658 == -0.0163
[2] 0.358 - 0.342 ==  0.0163

── 7. Error: we can separate struct columns (#690) (@test-column-extraction.R#92)  ────────────────────────────────────
cannot handle Array of type struct
1: sliding_window_sdf %>% sdf_separate_column("sw") at testthat/test-column-extraction.R:92
2: withVisible(eval(quote(`_fseq`(`_lhs`)), env, env))
3: eval(quote(`_fseq`(`_lhs`)), env, env)
4: eval(quote(`_fseq`(`_lhs`)), env, env)
5: `_fseq`(`_lhs`)
6: freduce(value, `_function_list`)
7: withVisible(function_list[[k]](value))
8: function_list[[k]](value)
9: sdf_separate_column(., "sw")
10: x %>% head(1) %>% dplyr::pull(!!rlang::sym(column)) %>% rlang::flatten() %>% length() %>% seq_len(.) at /Users/javierluraschi/RStudio/sparklyr/R/sdf_wrapper.R:293
...
32: withVisible(eval(quote(`_fseq`(`_lhs`)), env, env))
33: eval(quote(`_fseq`(`_lhs`)), env, env)
34: eval(quote(`_fseq`(`_lhs`)), env, env)
35: `_fseq`(`_lhs`)
36: freduce(value, `_function_list`)
37: function_list[[i]](value)
38: arrow_read_stream(.)
39: tibble::as_tibble(record_entry) at /Users/javierluraschi/RStudio/sparklyr/R/arrow_data.R:37
40: `as_tibble.arrow::RecordBatch`(record_entry)
41: RecordBatch__to_dataframe(x) at /Users/javierluraschi/RStudio/arrow/r/R/RecordBatch.R:59

── 8. Failure: top_n works as expected (@test-dplyr-top-n.R#23)  ──────────────────────────────────────────────────────
all(tn1 == tn2) isn't true.

── 9. Error: ml_recommend() works (@test-ml-als.R#44)  ────────────────────────────────────────────────────────────────
cannot handle Array of type struct
1: expect_identical(als_model %>% ml_recommend("users", 2) %>% colnames(), c("item", "recommendations", "user", "rating")) at testthat/test-ml-als.R:44
2: quasi_label(enquo(object), label)
3: eval_bare(get_expr(quo), get_env(quo))
4: als_model %>% ml_recommend("users", 2) %>% colnames()
5: withVisible(eval(quote(`_fseq`(`_lhs`)), env, env))
6: eval(quote(`_fseq`(`_lhs`)), env, env)
7: eval(quote(`_fseq`(`_lhs`)), env, env)
8: `_fseq`(`_lhs`)
9: freduce(value, `_function_list`)
10: function_list[[i]](value)
...
43: withVisible(eval(quote(`_fseq`(`_lhs`)), env, env))
44: eval(quote(`_fseq`(`_lhs`)), env, env)
45: eval(quote(`_fseq`(`_lhs`)), env, env)
46: `_fseq`(`_lhs`)
47: freduce(value, `_function_list`)
48: function_list[[i]](value)
49: arrow_read_stream(.)
50: tibble::as_tibble(record_entry) at /Users/javierluraschi/RStudio/sparklyr/R/arrow_data.R:37
51: `as_tibble.arrow::RecordBatch`(record_entry)
52: RecordBatch__to_dataframe(x) at /Users/javierluraschi/RStudio/arrow/r/R/RecordBatch.R:59

── 10. Failure: ml_bisecting_kmeans() works properly (@test-ml-clustering-bisecting-kmeans.R#255)  ────────────────────
`print(ml_bisecting_kmeans(iris_tbl, ~. - Species, k = 5, seed = 11))` has changed from known value recorded in 'output/print/bisecting-kmeans.txt'.
3/11 mismatches
x[5]: "1     4.803226    3.225806     1.419355   0.2096774"
y[5]: "1     4.750000    3.012500     1.666667   0.3166667"

x[6]: "2     5.290909    3.572727     1.759091   0.4045455"
y[6]: "2     5.217241    3.665517     1.472414   0.2689655"

x[11]: "Within Set Sum of Squared Errors =  63.01642"
y[11]: "Within Set Sum of Squared Errors =  60.60717"

── 11. Failure: we can construct a simple pivot table (@test-pivot.R#22)  ─────────────────────────────────────────────
unname(s) not equal to unname(r).
Component 2: Attributes: < Modes: list, NULL >
Component 2: Attributes: < Lengths: 1, 0 >
Component 2: Attributes: < names for target but not for current >
Component 2: Attributes: < current is not list-like >
Component 2: target is integer64, current is numeric
Component 3: Attributes: < Modes: list, NULL >
Component 3: Attributes: < Lengths: 1, 0 >
Component 3: Attributes: < names for target but not for current >
Component 3: Attributes: < current is not list-like >
...

── 12. Error: spark_read_json() can load data using column types (@test-read-write.R#49)  ─────────────────────────────
cannot handle Array of type struct
1: spark_read_json(sc, name = "iris_json_typed", path = "test.json", columns = list(Sepal_Length = "character", Species = "character", 
       Other = "struct<a:integer,b:character>")) %>% collect() at testthat/test-read-write.R:49
2: withVisible(eval(quote(`_fseq`(`_lhs`)), env, env))
3: eval(quote(`_fseq`(`_lhs`)), env, env)
4: eval(quote(`_fseq`(`_lhs`)), env, env)
5: `_fseq`(`_lhs`)
6: freduce(value, `_function_list`)
7: withVisible(function_list[[k]](value))
8: function_list[[k]](value)
9: collect(.)
10: collect.tbl_sql(.)
...
22: withVisible(eval(quote(`_fseq`(`_lhs`)), env, env))
23: eval(quote(`_fseq`(`_lhs`)), env, env)
24: eval(quote(`_fseq`(`_lhs`)), env, env)
25: `_fseq`(`_lhs`)
26: freduce(value, `_function_list`)
27: function_list[[i]](value)
28: arrow_read_stream(.)
29: tibble::as_tibble(record_entry) at /Users/javierluraschi/RStudio/sparklyr/R/arrow_data.R:37
30: `as_tibble.arrow::RecordBatch`(record_entry)
31: RecordBatch__to_dataframe(x) at /Users/javierluraschi/RStudio/arrow/r/R/RecordBatch.R:59

── 13. Error: (unknown) (@test-serialization.R#7)  ────────────────────────────────────────────────────────────────────
org.apache.spark.SparkException: Job aborted due to stage failure: Task 7 in stage 25364.0 failed 1 times, most recent failure: Lost task 7.0 in stage 25364.0 (TID 132972, localhost, executor driver): java.lang.UnsupportedOperationException: Unsupported data type: Date(MILLISECOND)

── 14. Error: 'spark_apply' can add columns (@test-spark-apply.R#28)  ─────────────────────────────────────────────────
cannot handle Array of type decimal
1: expect_equal(iris_tbl %>% spark_apply(function(e) cbind(e, 1), names = c(colnames(iris_tbl), "new")) %>% collect(), iris_tbl

── 15. Error: 'spark_apply' can roundtrip Date-Time (@test-spark-apply.R#188)  ────────────────────────────────────────
sparklyr worker rscript failure, check worker logs for details
    Log: /var/folders/ks/wm_bx4cn70s6h0r5vgqpsldm0000gn/T//RtmpOXR9Td/file1da775bc76cd_spark.log
18/10/25 15:36:32 ERROR sparklyr: RScript (7555) terminated unexpectedly: java.lang.UnsupportedOperationException: Unsupported data type: Date(MILLISECOND)

── 16. Error: 'spark_apply' supports grouped empty results (@test-spark-apply.R#215)  ─────────────────────────────────
sparklyr worker rscript failure, check worker logs for details
    Log: /var/folders/ks/wm_bx4cn70s6h0r5vgqpsldm0000gn/T//RtmpOXR9Td/file1da775bc76cd_spark.log
18/10/25 15:36:44 ERROR sparklyr: RScript (3929) terminated unexpectedly: java.util.NoSuchElementException
	at org.apache.spark.sql.vectorized.ColumnarBatch$1.next(ColumnarBatch.java:65)

── 17. Failure: debug_string works (@test-spark-utils.R#38)  ──────────────────────────────────────────────────────────
grepl("^\\(1\\)", debug[1]) isn't true.

[javierluraschi]

Merged into #1611