max() does not work on strings, head() function todo about rumble HOT 10 CLOSED

Here there is a bug, because max should also work on sequences of strings.

from rumble.

Then there are two bugs: the one with max on strings and the bug that I reported.

When you run the first of the two queries, Sparksoniq enters into an invalid state, in which no query cannot be evaluated anymore.

from rumble.

To summarize, there are thus three bugs:

• max with strings
• implement the head function
• investigate why queries stop working after an invalid state is reached

from rumble.

Note: head() should work like count() as recently implemented by Can, i.e., if the parameter is an RDD, it should invoke first(). If it is not, it should materialize the input and count locally.

from rumble.

I have tried recreating the mentioned problem with the given query in my local environment with the sample json dataset. When I ran the two queries shown in bold below in succession, the program didn't enter an invalid state and results were returned correctly. Do you have any suggestions on what might be causing the invalid state or how I can trigger it?

jiqs$ for $o in json-file("./src/main/resources/queries/confusion_sample.json")
where $o.date eq max(for $o in json-file("./src/main/resources/queries/confusion_sample.json") return $o.date)
return $o

[ERROR] Job aborted due to stage failure: Task 0 in stage 0.0 failed 1 times, most recent failure: Lost task 0.0 in stage 0.0 (TID 0, localhost, executor driver): sparksoniq.exceptions.UnexpectedTypeException: Error [err: XPTY0004]LINE:2:COLUMN:17:Max expression has non numeric args 2013-08-19

jiqs$ for $o in json-file("./src/main/resources/queries/confusion_sample.json")
where $o.guess eq "Dinka"
return $o
{ "date" : "2013-08-19", "country" : "AU", "guess" : "Dinka", "choices" : [ Danish, Dinka, Khmer, Lao ], "sample" : "af5e8f27cef9e689a070b8814dcc02c3", "target" : "Dinka" }
{ "date" : "2013-08-19", "country" : "AU", "guess" : "Dinka", "choices" : [ Amharic, Dinka, Mandarin, Yiddish ], "sample" : "cd3dffdb83f221e09710cbfbcaffdb31", "target" : "Dinka" }
{ "date" : "2013-08-19", "country" : "AU", "guess" : "Dinka", "choices" : [ Amharic, Arabic, Dinka, Samoan ], "sample" : "509c36eb58dbce009ccf93f375358d53", "target" : "Samoan" }
{ "date" : "2013-08-20", "country" : "AU", "guess" : "Dinka", "choices" : [ Dinka, Korean, Swedish, Thai ], "sample" : "0f8e6818d2be77b75712bc9ea4e87171", "target" : "Korean" }
{ "date" : "2013-08-20", "country" : "AU", "guess" : "Dinka", "choices" : [ Dinka, Malayalam, Sinhalese, Ukrainian ], "sample" : "73a09a8879e70e8648d7d75d656a5b79", "target" : "Malayalam" }
{ "date" : "2013-08-20", "country" : "AU", "guess" : "Dinka", "choices" : [ Dinka, Malayalam, Tamil, Urdu ], "sample" : "cc305dcad696147e28c6f36615aa0952", "target" : "Tamil" }
{ "date" : "2013-08-20", "country" : "AU", "guess" : "Dinka", "choices" : [ Bulgarian, Dinka, French, Khmer ], "sample" : "69a5c674fe40942aaae4d10c9df46ee0", "target" : "Bulgarian" }

from rumble.

Max with strings implemented.
head function implemented (without RDD support, issue opened for this)
I couldn't recreate the invalid state with the queries I've given above. Any tips on how they can be reached or can we close this issue?

from rumble.

I think the invalid state only happens when a query has a nested FLWOR. I would simply close because we will take care of it later.

from rumble.

I cannot reproduce it myself anymore with the 0.9.4 release.

from rumble.

I tried version 0.9.2 as well and cannot reproduce it there either. Sorry that I didn't document better how to reproduce it.

from rumble.

@CanBerker I think we should go ahead and merge.

from rumble.