June 2015 - users-list

MonetDB performance disappointing
by Martin Schwitalla 30 Jun '15

30 Jun '15

Hello, I`m doing some experiments with a database, which originally runs on PostgreSQL. But some queries take very long on this system. Especially one query type, which takes about 14s to finish. The idea was to port it to MonetDB, because the database is not that big (~8GB) and it`s an analytical task. But suprisingly MonetDB takes even longer to finish. With the default pipeline it takes about 22-23s. I tried various pipelines and the minimal_pipe was the best. It finished the task in about 16,5s, better, but still not great. I`m asking myself what I could do to further improve the performance. I added some primary and foreign key constraints and set the database to readonly, but it didn`t do much. The trace tells me that some joins take the most time and some subselects also take much time. But nothing besides that. The query has 10 joins. Do you have any advices for me? Best regards, Martin

5 11

Concurrency test causes crash of MonetDB 5 server
by michael.witt＠uniserv.com 30 Jun '15

30 Jun '15

Hi, we experience a crash of MonetDB 5 server during tests of a multi-threaded client access via JDBC-Driver and MapiSocket: The application is already optimized to avoid "concurrency conflicts" because of the "Optimistic Concurrency Control" concept. The test simulates client access via two threads. All update operations (INSERT, UPDATE, DELETE, DROP TABLE, CREATE TABLE) are done in serial. All read operations are done in parallel to other read operations or to update operations. The test causes a crash of mserver5.exe after a few minutes. If the test uses just one thread, all read and update operations are done in serial and the server does not crash. What is the rule of thumb to avoid crashes when using several client threads? Crash details: Problem signature: Problem Event Name: APPCRASH Application Name: mserver5.exe Application Version: 0.0.0.0 Application Timestamp: 55389f58 Fault Module Name: StackHash_8ec6 Fault Module Version: 6.1.7601.18247 Fault Module Timestamp: 521eaf24 Exception Code: c0000374 Exception Offset: 00000000000c4102 OS Version: 6.1.7601.2.1.0.144.8 Locale ID: 1031 Additional Information 1: 8ec6 Additional Information 2: 8ec62499039d1d8234b025a62c725eb7 Additional Information 3: f8f3 Additional Information 4: f8f3e520847c0684b15d4080171779ef # MonetDB 5 server v11.19.9 "Oct2014-SP2" # Serving database 'uniserv', using 4 threads # Compiled for x86_64-pc-winnt/64bit with 64bit OIDs dynamically linked # Found 10.000 GiB available main-memory. .. # MonetDB/SQL module loaded Michael Witt

2 2

JDBC PreparedStatement crashing the Database!!
by Vijay Krishna 29 Jun '15

29 Jun '15

Hi, I recently faced an issue with MonetDB JDBC PreparedStatement. I have a table with 15 columns out of which 14 are foreign keys to columns in another 14 tables. When I try to insert a valid record (that satisfies all the FK relations) using PreparedStatement, the database gets crashed while executing the line, preparedStatement = connection.prepareStatement(sqlString); On trial and error testing, I found that this works very fine until a table has 13 FK relations. If the tables has 14 FKs, then the database crashes. On a brighter side, this operation works very fine in terminal (mclient) with 14 FKs even. Is there any limit on the FKs with prepared statement? Or what am I doing wrong here? Any help much appreciated. Thanks in advance. Regards, Vijayakrishna.P. Mobile : (+91) 9500402305.

2 2

Is minimal_pipe enforced on long MAL plans?
by Roberto Cornacchia 22 Jun '15

22 Jun '15

Hi there, Is minimal_pipe enforced under any condition? For example when the the MAL plan generated by rel2bin is longer than a given threshold? Context: I have a pretty long query which results in a MAL plan of 33K lines. Though the query is a bit long, I still would expect a 2-3K lines MAL plan. The plan indeed contains the same pieces of code re-evaluated over and over. It first, I simply blamed commonTerms not doing its job. The thing is, the MAL plans generated with minimal_pipe and default_pipe are absolutely identical for this query. Hence my question above. Best, Roberto

3 3

MonetDb Udf
by Fatma Nur Özdemir 19 Jun '15

19 Jun '15

Hello, We are using MonetDB5 server v11.19.15 “Oct2014-SP4” windows version and we want to use a function that we wrote in C programming language. However, we cannot use this user defined function within MonetDB. Maybe, this problem is about windows version because we are not sure that windows version of MonetDb supports this feature. If anybody knows something about this, please inform us. Thank you in advance. Fatma Nur [http://www.bilyoner.com/fm/mail/mailalti_Bmw-applewatch.jpg]<http://www.bilyoner.com/kampanyalar/2015haziranagustosmp.jspx> Bu elektronik posta ve onunla iletilen butun dosyalar sadece gondericisi tarafindan almasi amaclanan yetkili gercek ya da tuzel kisinin kullanimi icindir. Eger soz konusu yetkili alici degilseniz bu elektronik postanin icerigini aciklamaniz, kopyalamaniz, yonlendirmeniz ve kullanmaniz hukuka aykiri ve yasaktir, eger bu mesaji yanlislikla aldiysaniz lutfen gondereni ile temasa geciniz ve bu elektronik postayi siliniz. B�LYONER �nteraktif Hizmetleri A.S. bu mesajin icerdigi bilgilerin dogrulugu veya eksiksiz oldugu konusunda herhangi bir garanti vermemektedir. Bu nedenle bu bilgilerin ne sekilde olursa olsun iceriginden, iletilmesinden, alinmasindan ve saklanmasindan sorumlu degildir. Bu mesajdaki gorusler yalnizca gonderen kisiye aittir ve Bilyoner �nteraktif Hizmetleri A.S. nin goruslerini yansitmayabilir. Bu e-posta bilinen butun bilgisayar viruslerine karsi taranmistir.

1 0

Re: Monetdb-R integration
by imad hajj chahine 19 Jun '15

19 Jun '15

Hi Hannes, This error is due that the "2-age" column has more than 2 distinct values, there is a space after some values. can you run a trim on the "2-Age" and then try it again. As for the vector thing, I didnt get the point, shouldnt the db engine split the data into vectors and feed the aggregate function which return one data point? Thanks On Fri, Jun 19, 2015 at 11:12 AM, Hannes Mühleisen <Hannes.Muehleisen(a)cwi.nl > wrote: > I am running attached script, which I pieced together from your emails. > This does not run, it produces the following error: > > Error running R expression. Error message: Error in t.test.formula(arg1 ~ > arg2) : > grouping factor must have exactly 2 levels > Calls: as.data.frame -> <Anonymous> -> t.test -> t.test.formula > > > > > In general, for aggregations, the R code needs to return a vector with a > single entry for each group. > > > Best, > > Hannes > > > > > On 19 Jun 2015, at 09:44, imad hajj chahine <imad.hajj.chahine(a)gmail.com> > wrote: > > > > Thanks, if you need more information from my side let me know. > > > > Also can you answer me on the following when i use 2 functions one to > retrieve Pvalue and one for the stat value: > > Am I doing the correct way to retrieve the stat and p-value, is the > ttest get executed twice in this case? is there a more optimal way where i > can retrieve both information. > > > > On Fri, Jun 19, 2015 at 10:39 AM, Hannes Mühleisen < > Hannes.Muehleisen(a)cwi.nl> wrote: > > Looking into this at the moment. > > > > > On 16 Jun 2015, at 09:56, Dimitar Nedev < > dimitar.nedev(a)monetdbsolutions.com> wrote: > > > > > > Hannes, do you know if the STD_DEV is the issue there or if the > queries should actually yield different results? > > > > > > Dimitar > > > > > >> On 2015-Jun-15, at 13:28 , imad hajj chahine < > imad.hajj.chahine(a)gmail.com> wrote: > > >> > > >> Hi, > > >> > > >> Shouldn't we have the same result from running the first query and > the second query? > > >> after all its an aggregate function and should return the test for > each group. > > >> > > >> I checked that monetdb had problem with running group by with > statistics functions like STD_DEV, are these 2 issues related? > > >> > > >> Thank you. > > >> > > >> > > >> > > >> > > >> On Mon, Jun 15, 2015 at 1:52 PM, Dimitar Nedev < > dimitar.nedev(a)monetdbsolutions.com> wrote: > > >> Hi Imad, > > >> > > >> Here is the output I got from running the queries. Note, this might > not be entirely representative since I did clean-up some data before > loading it. > > >> > > >> sql>select "2-Age", "tone"."ttest2"("9-Score",6.08963) from > "tone"."Marketing_Loyalty_4700298d-9862-40b3-9028-b0f15dab9dea" where > "2-Age" is not null group by "2-Age"; > > >> +---------+--------------------------+ > > >> | 2-Age | L1 | > > >> +=========+==========================+ > > >> | [10-20] | 0.89311103869734754 | > > >> +---------+--------------------------+ > > >> 3 tuples (3.218ms) > > >> > > >> sql>select '[10-20]' as "2-Age","tone"."ttest2"("9-Score",6.08963) > from "tone"."Marketing_Loyalty_4700298d-9862-40b3-9028-b0f15dab9dea" where > "2-Age" = '[10-20]' > > >> more>union all > > >> more>select '[21-30]' as "2-Age","tone"."ttest2"("9-Score",6.08963) > from "tone"."Marketing_Loyalty_4700298d-9862-40b3-9028-b0f15dab9dea" where > "2-Age" = '[21-30]' > > >> more>union all > > >> more>select '[> 30]' as "2-Age","tone"."ttest2"("9-Score",6.08963) > from "tone"."Marketing_Loyalty_4700298d-9862-40b3-9028-b0f15dab9dea" where > "2-Age" = '> 30'; > > >> +---------+--------------------------+ > > >> | 2-Age | L1 | > > >> +=========+==========================+ > > >> | [10-20] | 0.14150742355349472 | > > >> | [21-30] | 0.32872830865003566 | > > >> | [> 30] | 0.37127075363219841 | > > >> +---------+--------------------------+ > > >> 3 tuples (8.825ms) > > >> > > >> > > >> sql>select "tone"."ttest2samplesStatistic"("9-Score", "4-Gender") as > "stat", "tone"."ttest2samples"("9-Score", "4-Gender") as "pvalue" from > "tone"."Marketing_Loyalty_4700298d-9862-40b3-9028-b0f15dab9dea" group by > "2-Age"; > > >> +--------------------------+--------------------------+ > > >> | stat | pvalue | > > >> +==========================+==========================+ > > >> | -2.1588506639013736 | 0.032555033970368859 | > > >> +--------------------------+--------------------------+ > > >> 1 tuple (13.287ms) > > >> > > >> > > >> I dropped the aggregate with: > > >> sql>DROP AGGREGATE "tone"."ttest2samplesStatistic"; > > >> operation successful (2.049ms) > > >> > > >> > > >> Best regards, > > >> Dimitar > > >> > > >>> On 2015-Jun-12, at 00:46 , imad hajj chahine < > imad.hajj.chahine(a)gmail.com> wrote: > > >>> > > >>> Thank you Dimitar, > > >>> > > >>> I was using Squirrel to connect from my host machine to the VM. > > >>> > > >>> Now when I connect using SSH-client and run the short group by query > it works but the result is not similar to the union queries (PS: i also run > it in the VM direclty and had the same result) > > >>> > > >>> > > >>> +---------+--------------------------+ > > >>> | 2-Age | pval | > > >>> +=========+==========================+ > > >>> | [21-30] | 0.99996611809271041 | > > >>> +---------+--------------------------+ > > >>> 3 tuples (2.638ms) > > >>> > > >>> > > >>> What i am trying to achieve is to run a one sample ttest for all the > age categories, and it seems that the group by is ignored. > > >>> > > >>> I have also created 2 functions for the two sample ttest: > > >>> > > >>> CREATE AGGREGATE "tone"."ttest2samples"(arg1 double, arg2 > varchar(250)) RETURNS double LANGUAGE R { t.test(arg1 ~ arg2)$p.value }; > > >>> CREATE AGGREGATE "tone"."ttest2samplesStatistic"(arg1 double, arg2 > varchar(250)) RETURNS double LANGUAGE R { t.test(arg1 ~ arg2)$statistic }; > > >>> > > >>> select "tone"."ttest2samplesStatistic"("9-Score", "4-Gender") as > "stat", "tone"."ttest2samples"("9-Score", "4-Gender") as "pvalue" from > "tone"."Marketing_Loyalty_4700298d-9862-40b3-9028-b0f15dab9dea" group by > "2-Age" > > >>> > > >>> Am I doing the correct way to retrieve the stat and p-value, is the > ttest get executed twice in this case? is there a more optimal way where i > can retrieve both information. Please note that I also have the same > grouping issue as before. > > >>> > > >>> Do you have the same behavior, or I am missing something in the > creation of these functions. > > >>> > > >>> A final question, how do u drop a R function? i am trying to execute > the following: > > >>> drop all function "tone"."ttest2samplesStat" > > >>> > > >>> > > >>> Thank you. > > >>> > > >>> > > >>> > > >>> On Thu, Jun 11, 2015 at 11:28 PM, Dimitar Nedev < > dimitar.nedev(a)monetdbsolutions.com> wrote: > > >>> Hi Imad, > > >>> > > >>> Hannes asked me to look at your issue with the MonetDB-R VM. > > >>> > > >>> I loaded the data and ran the queries you mentioned, but I did not > encounter any issues. This holds for both the short and the union queries. > I am also not familiar with the issue you report, nor its cause. At this > time I think it is local to your instance. If you can export your db, or > re-load the data, I would recommend you to create a new instance with a > fresh image and try again. If you import the image as a new VM instance, > you can keep your old one around as well. > > >>> > > >>> Please note that during the test: > > >>> - I ran as the admin (monetdb) user, as the one created default on > the VM image cannot create new schemas. > > >>> - I have to clean the data a bit up to load it, removing dozen lines > or so. > > >>> - Renamed ttest1 to ttest2 to match the queries. > > >>> > > >>> Best regards, > > >>> Dimitar > > >>> > > >>>>>>>>> Begin forwarded message: > > >>>>>>>>> > > >>>>>>>>> From: imad hajj chahine <imad.hajj.chahine(a)gmail.com> > > >>>>>>>>> Subject: Re: Monetdb-R integration > > >>>>>>>>> Date: 10 Jun 2015 10:11:48 CEST > > >>>>>>>>> To: Hannes Mühleisen <Hannes.Muehleisen(a)cwi.nl> > > >>>>>>>>> > > >>>>>>>>> Hi, > > >>>>>>>>> > > >>>>>>>>> I am using the virtual box provided from > https://www.monetdbsolutions.com/solutions/cloud/vm: > > >>>>>>>>> OS:linux version 3.10.0-123.9.3.el7.x86_64 ( > builder(a)kbuilder.dev.centos.org) > > >>>>>>>>> Monetdb: V1.7: (Oct2014-SP1) > > >>>>>>>>> R: 3.1.2 (2014-10-31) > > >>>>>>>>> > > >>>>>>>>> > > >>>>>>>>> > > >>>>>>>>> create table > "tone"."Marketing_Loyalty_4700298d-9862-40b3-9028-b0f15dab9dea" ("Consumer > Number" INTEGER,"1-Disloyalty" varchar(250),"2-Age" varchar(250),"3-Amount > transferred" INTEGER,"4-Gender" varchar(250),"5-Citiz enship" > varchar(250),"6- Profession" varchar(250),"7-City" varchar(250),"8-1 NC > Billboard" varchar(250),"8-2 NC Friends" varchar(250),"8-3 NC Magazine" > varchar(250),"8-4 NC Others" varchar(250),"9-Score" decimal(18,3),"10- Type > of operation" varchar(250),"11-1 Welcoming" INTEGER,"11-2 Waiting time" > INTEGER,"11-3 Time to finish request" INTEGER,"11-4 Fees" INTEGER,"11-5 > Courtesy of employees" INTEGER,"11-6 Branch layout" INTEGER,"11-7 Attention > to special requests" INTEGER,"12- Total time for service" > decimal(18,3),"13- Distance from POS" decimal(18,3),"14- Weight" INTEGER) > > >>>>>>>>> > > >>>>>>>>> COPY offset 2 INTO > "tone"."Marketing_Loyalty_4700298d-9862-40b3-9028-b0f15dab9dea" FROM > '/home/App_Data/TestingDataSimple_2003_Marketing_Loyalty_4700298d-9862-40b3-9028-b0f15dab9dea.csv' > USING DELIMITERS ',','\n','"'; > > >>>>>>>>> > > >>>>>>>>> Thanks > > >>>>>>>>> > > >>>>>>>>> On Wed, Jun 10, 2015 at 10:50 AM, Hannes Mühleisen < > Hannes.Muehleisen(a)cwi.nl> wrote: > > >>>>>>>>> Hello Imad, > > >>>>>>>>> > > >>>>>>>>> which version of MonetDB and R are you using? Which operating > system? > > >>>>>>>>> > > >>>>>>>>> Can you please also include the commands you used to load your > CSV into MonetDB/ create the table? > > >>>>>>>>> > > >>>>>>>>> Thanks, > > >>>>>>>>> > > >>>>>>>>> Hannes > > >>>>>>>>> > > >>>>>>>>>> On 10 Jun 2015, at 00:25, imad hajj chahine < > imad.hajj.chahine(a)gmail.com> wrote: > > >>>>>>>>>> > > >>>>>>>>>> Hi Hannes, > > >>>>>>>>>> > > >>>>>>>>>> I am testing Monetdb-R integration to do some statistical > test, i followed the tutorial you posted on youtube and i have created the > following function: > > >>>>>>>>>> > > >>>>>>>>>> CREATE AGGREGATE "tone"."ttest1"(arg1 double, arg2 double) > RETURNS double LANGUAGE R { t.test(arg1, arg2)$p.value }; > > >>>>>>>>>> > > >>>>>>>>>> this function will basically do a one sample t-test. > > >>>>>>>>>> > > >>>>>>>>>> I have loaded the sample attached file into a demo database > and trying to run the following query: > > >>>>>>>>>> > > >>>>>>>>>> select "2-Age", "tone"."ttest2"("9-Score",6.08963) from > "tone"."Marketing_Loyalty_4700298d-9862-40b3-9028-b0f15dab9dea" where > "2-Age" is not null group by "2-Age" > > >>>>>>>>>> > > >>>>>>>>>> The following error happened > > >>>>>>>>>> Error: protocol violation: unexpected line in data block: . > > >>>>>>>>>> SQLState: M0M10 > > >>>>>>>>>> ErrorCode: 0 > > >>>>>>>>>> > > >>>>>>>>>> alternatively i managed to have the intended result by > running this query: > > >>>>>>>>>> > > >>>>>>>>>> select '[10-20]' as > "2-Age","tone"."ttest2"("9-Score",6.08963) from > "tone"."Marketing_Loyalty_4700298d-9862-40b3-9028-b0f15dab9dea" where > "2-Age" = '[10-20]' > > >>>>>>>>>> union all > > >>>>>>>>>> select '[21-30]' as > "2-Age","tone"."ttest2"("9-Score",6.08963) from > "tone"."Marketing_Loyalty_4700298d-9862-40b3-9028-b0f15dab9dea" where > "2-Age" = '[21-30]' > > >>>>>>>>>> union all > > >>>>>>>>>> select '[> 30]' as "2-Age","tone"."ttest2"("9-Score",6.08963) > from "tone"."Marketing_Loyalty_4700298d-9862-40b3-9028-b0f15dab9dea" where > "2-Age" = '> 30' > > >>>>>>>>>> > > >>>>>>>>>> > > >>>>>>>>>> any help or guidance is appreciated. > > >>>>>>>>>> Thank you. > > >>>>>>>>>> > > >>>>>>>>>> > <TestingDataSimple_2003_Marketing_Loyalty_4700298d-9862-40b3-9028-b0f15dab9dea.csv> > > >>>>>>>>> > > >>>>>>>>> > > >>>>>>>> > > >>>>>>> > > >>>>>> > > >>>>> > > >>>> > > >>> > > >> > > >> > > > > > > > > > >

1 0

Fwd: Monetdb-R integration
by Hannes Mühleisen 19 Jun '15

19 Jun '15

(Forwarding to users list, free support is public) > Begin forwarded message: > > From: Hannes Mühleisen <hannes.muehleisen(a)cwi.nl> > Subject: Re: Monetdb-R integration > Date: 19 Jun 2015 10:12:48 CEST > To: imad hajj chahine <imad.hajj.chahine(a)gmail.com> > Cc: Dimitar Nedev <dimitar.nedev(a)monetdbsolutions.com> > > I am running attached script, which I pieced together from your emails. This does not run, it produces the following error: > > Error running R expression. Error message: Error in t.test.formula(arg1 ~ arg2) : > grouping factor must have exactly 2 levels > Calls: as.data.frame -> <Anonymous> -> t.test -> t.test.formula > > > > In general, for aggregations, the R code needs to return a vector with a single entry for each group. > > > Best, > > Hannes > > > >> On 19 Jun 2015, at 09:44, imad hajj chahine <imad.hajj.chahine(a)gmail.com> wrote: >> >> Thanks, if you need more information from my side let me know. >> >> Also can you answer me on the following when i use 2 functions one to retrieve Pvalue and one for the stat value: >> Am I doing the correct way to retrieve the stat and p-value, is the ttest get executed twice in this case? is there a more optimal way where i can retrieve both information. >> >> On Fri, Jun 19, 2015 at 10:39 AM, Hannes Mühleisen <Hannes.Muehleisen(a)cwi.nl> wrote: >> Looking into this at the moment. >> >>> On 16 Jun 2015, at 09:56, Dimitar Nedev <dimitar.nedev(a)monetdbsolutions.com> wrote: >>> >>> Hannes, do you know if the STD_DEV is the issue there or if the queries should actually yield different results? >>> >>> Dimitar >>> >>>> On 2015-Jun-15, at 13:28 , imad hajj chahine <imad.hajj.chahine(a)gmail.com> wrote: >>>> >>>> Hi, >>>> >>>> Shouldn't we have the same result from running the first query and the second query? >>>> after all its an aggregate function and should return the test for each group. >>>> >>>> I checked that monetdb had problem with running group by with statistics functions like STD_DEV, are these 2 issues related? >>>> >>>> Thank you. >>>> >>>> >>>> >>>> >>>> On Mon, Jun 15, 2015 at 1:52 PM, Dimitar Nedev <dimitar.nedev(a)monetdbsolutions.com> wrote: >>>> Hi Imad, >>>> >>>> Here is the output I got from running the queries. Note, this might not be entirely representative since I did clean-up some data before loading it. >>>> >>>> sql>select "2-Age", "tone"."ttest2"("9-Score",6.08963) from "tone"."Marketing_Loyalty_4700298d-9862-40b3-9028-b0f15dab9dea" where "2-Age" is not null group by "2-Age"; >>>> +---------+--------------------------+ >>>> | 2-Age | L1 | >>>> +=========+==========================+ >>>> | [10-20] | 0.89311103869734754 | >>>> +---------+--------------------------+ >>>> 3 tuples (3.218ms) >>>> >>>> sql>select '[10-20]' as "2-Age","tone"."ttest2"("9-Score",6.08963) from "tone"."Marketing_Loyalty_4700298d-9862-40b3-9028-b0f15dab9dea" where "2-Age" = '[10-20]' >>>> more>union all >>>> more>select '[21-30]' as "2-Age","tone"."ttest2"("9-Score",6.08963) from "tone"."Marketing_Loyalty_4700298d-9862-40b3-9028-b0f15dab9dea" where "2-Age" = '[21-30]' >>>> more>union all >>>> more>select '[> 30]' as "2-Age","tone"."ttest2"("9-Score",6.08963) from "tone"."Marketing_Loyalty_4700298d-9862-40b3-9028-b0f15dab9dea" where "2-Age" = '> 30'; >>>> +---------+--------------------------+ >>>> | 2-Age | L1 | >>>> +=========+==========================+ >>>> | [10-20] | 0.14150742355349472 | >>>> | [21-30] | 0.32872830865003566 | >>>> | [> 30] | 0.37127075363219841 | >>>> +---------+--------------------------+ >>>> 3 tuples (8.825ms) >>>> >>>> >>>> sql>select "tone"."ttest2samplesStatistic"("9-Score", "4-Gender") as "stat", "tone"."ttest2samples"("9-Score", "4-Gender") as "pvalue" from "tone"."Marketing_Loyalty_4700298d-9862-40b3-9028-b0f15dab9dea" group by "2-Age"; >>>> +--------------------------+--------------------------+ >>>> | stat | pvalue | >>>> +==========================+==========================+ >>>> | -2.1588506639013736 | 0.032555033970368859 | >>>> +--------------------------+--------------------------+ >>>> 1 tuple (13.287ms) >>>> >>>> >>>> I dropped the aggregate with: >>>> sql>DROP AGGREGATE "tone"."ttest2samplesStatistic"; >>>> operation successful (2.049ms) >>>> >>>> >>>> Best regards, >>>> Dimitar >>>> >>>>> On 2015-Jun-12, at 00:46 , imad hajj chahine <imad.hajj.chahine(a)gmail.com> wrote: >>>>> >>>>> Thank you Dimitar, >>>>> >>>>> I was using Squirrel to connect from my host machine to the VM. >>>>> >>>>> Now when I connect using SSH-client and run the short group by query it works but the result is not similar to the union queries (PS: i also run it in the VM direclty and had the same result) >>>>> >>>>> >>>>> +---------+--------------------------+ >>>>> | 2-Age | pval | >>>>> +=========+==========================+ >>>>> | [21-30] | 0.99996611809271041 | >>>>> +---------+--------------------------+ >>>>> 3 tuples (2.638ms) >>>>> >>>>> >>>>> What i am trying to achieve is to run a one sample ttest for all the age categories, and it seems that the group by is ignored. >>>>> >>>>> I have also created 2 functions for the two sample ttest: >>>>> >>>>> CREATE AGGREGATE "tone"."ttest2samples"(arg1 double, arg2 varchar(250)) RETURNS double LANGUAGE R { t.test(arg1 ~ arg2)$p.value }; >>>>> CREATE AGGREGATE "tone"."ttest2samplesStatistic"(arg1 double, arg2 varchar(250)) RETURNS double LANGUAGE R { t.test(arg1 ~ arg2)$statistic }; >>>>> >>>>> select "tone"."ttest2samplesStatistic"("9-Score", "4-Gender") as "stat", "tone"."ttest2samples"("9-Score", "4-Gender") as "pvalue" from "tone"."Marketing_Loyalty_4700298d-9862-40b3-9028-b0f15dab9dea" group by "2-Age" >>>>> >>>>> Am I doing the correct way to retrieve the stat and p-value, is the ttest get executed twice in this case? is there a more optimal way where i can retrieve both information. Please note that I also have the same grouping issue as before. >>>>> >>>>> Do you have the same behavior, or I am missing something in the creation of these functions. >>>>> >>>>> A final question, how do u drop a R function? i am trying to execute the following: >>>>> drop all function "tone"."ttest2samplesStat" >>>>> >>>>> >>>>> Thank you. >>>>> >>>>> >>>>> >>>>> On Thu, Jun 11, 2015 at 11:28 PM, Dimitar Nedev <dimitar.nedev(a)monetdbsolutions.com> wrote: >>>>> Hi Imad, >>>>> >>>>> Hannes asked me to look at your issue with the MonetDB-R VM. >>>>> >>>>> I loaded the data and ran the queries you mentioned, but I did not encounter any issues. This holds for both the short and the union queries. I am also not familiar with the issue you report, nor its cause. At this time I think it is local to your instance. If you can export your db, or re-load the data, I would recommend you to create a new instance with a fresh image and try again. If you import the image as a new VM instance, you can keep your old one around as well. >>>>> >>>>> Please note that during the test: >>>>> - I ran as the admin (monetdb) user, as the one created default on the VM image cannot create new schemas. >>>>> - I have to clean the data a bit up to load it, removing dozen lines or so. >>>>> - Renamed ttest1 to ttest2 to match the queries. >>>>> >>>>> Best regards, >>>>> Dimitar >>>>> >>>>>>>>>>> Begin forwarded message: >>>>>>>>>>> >>>>>>>>>>> From: imad hajj chahine <imad.hajj.chahine(a)gmail.com> >>>>>>>>>>> Subject: Re: Monetdb-R integration >>>>>>>>>>> Date: 10 Jun 2015 10:11:48 CEST >>>>>>>>>>> To: Hannes Mühleisen <Hannes.Muehleisen(a)cwi.nl> >>>>>>>>>>> >>>>>>>>>>> Hi, >>>>>>>>>>> >>>>>>>>>>> I am using the virtual box provided from https://www.monetdbsolutions.com/solutions/cloud/vm: >>>>>>>>>>> OS:linux version 3.10.0-123.9.3.el7.x86_64 (builder(a)kbuilder.dev.centos.org) >>>>>>>>>>> Monetdb: V1.7: (Oct2014-SP1) >>>>>>>>>>> R: 3.1.2 (2014-10-31) >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> create table "tone"."Marketing_Loyalty_4700298d-9862-40b3-9028-b0f15dab9dea" ("Consumer Number" INTEGER,"1-Disloyalty" varchar(250),"2-Age" varchar(250),"3-Amount transferred" INTEGER,"4-Gender" varchar(250),"5-Citiz enship" varchar(250),"6- Profession" varchar(250),"7-City" varchar(250),"8-1 NC Billboard" varchar(250),"8-2 NC Friends" varchar(250),"8-3 NC Magazine" varchar(250),"8-4 NC Others" varchar(250),"9-Score" decimal(18,3),"10- Type of operation" varchar(250),"11-1 Welcoming" INTEGER,"11-2 Waiting time" INTEGER,"11-3 Time to finish request" INTEGER,"11-4 Fees" INTEGER,"11-5 Courtesy of employees" INTEGER,"11-6 Branch layout" INTEGER,"11-7 Attention to special requests" INTEGER,"12- Total time for service" decimal(18,3),"13- Distance from POS" decimal(18,3),"14- Weight" INTEGER) >>>>>>>>>>> >>>>>>>>>>> COPY offset 2 INTO "tone"."Marketing_Loyalty_4700298d-9862-40b3-9028-b0f15dab9dea" FROM '/home/App_Data/TestingDataSimple_2003_Marketing_Loyalty_4700298d-9862-40b3-9028-b0f15dab9dea.csv' USING DELIMITERS ',','\n','"'; >>>>>>>>>>> >>>>>>>>>>> Thanks >>>>>>>>>>> >>>>>>>>>>> On Wed, Jun 10, 2015 at 10:50 AM, Hannes Mühleisen <Hannes.Muehleisen(a)cwi.nl> wrote: >>>>>>>>>>> Hello Imad, >>>>>>>>>>> >>>>>>>>>>> which version of MonetDB and R are you using? Which operating system? >>>>>>>>>>> >>>>>>>>>>> Can you please also include the commands you used to load your CSV into MonetDB/ create the table? >>>>>>>>>>> >>>>>>>>>>> Thanks, >>>>>>>>>>> >>>>>>>>>>> Hannes >>>>>>>>>>> >>>>>>>>>>>> On 10 Jun 2015, at 00:25, imad hajj chahine <imad.hajj.chahine(a)gmail.com> wrote: >>>>>>>>>>>> >>>>>>>>>>>> Hi Hannes, >>>>>>>>>>>> >>>>>>>>>>>> I am testing Monetdb-R integration to do some statistical test, i followed the tutorial you posted on youtube and i have created the following function: >>>>>>>>>>>> >>>>>>>>>>>> CREATE AGGREGATE "tone"."ttest1"(arg1 double, arg2 double) RETURNS double LANGUAGE R { t.test(arg1, arg2)$p.value }; >>>>>>>>>>>> >>>>>>>>>>>> this function will basically do a one sample t-test. >>>>>>>>>>>> >>>>>>>>>>>> I have loaded the sample attached file into a demo database and trying to run the following query: >>>>>>>>>>>> >>>>>>>>>>>> select "2-Age", "tone"."ttest2"("9-Score",6.08963) from "tone"."Marketing_Loyalty_4700298d-9862-40b3-9028-b0f15dab9dea" where "2-Age" is not null group by "2-Age" >>>>>>>>>>>> >>>>>>>>>>>> The following error happened >>>>>>>>>>>> Error: protocol violation: unexpected line in data block: . >>>>>>>>>>>> SQLState: M0M10 >>>>>>>>>>>> ErrorCode: 0 >>>>>>>>>>>> >>>>>>>>>>>> alternatively i managed to have the intended result by running this query: >>>>>>>>>>>> >>>>>>>>>>>> select '[10-20]' as "2-Age","tone"."ttest2"("9-Score",6.08963) from "tone"."Marketing_Loyalty_4700298d-9862-40b3-9028-b0f15dab9dea" where "2-Age" = '[10-20]' >>>>>>>>>>>> union all >>>>>>>>>>>> select '[21-30]' as "2-Age","tone"."ttest2"("9-Score",6.08963) from "tone"."Marketing_Loyalty_4700298d-9862-40b3-9028-b0f15dab9dea" where "2-Age" = '[21-30]' >>>>>>>>>>>> union all >>>>>>>>>>>> select '[> 30]' as "2-Age","tone"."ttest2"("9-Score",6.08963) from "tone"."Marketing_Loyalty_4700298d-9862-40b3-9028-b0f15dab9dea" where "2-Age" = '> 30' >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>> any help or guidance is appreciated. >>>>>>>>>>>> Thank you. >>>>>>>>>>>> >>>>>>>>>>>> <TestingDataSimple_2003_Marketing_Loyalty_4700298d-9862-40b3-9028-b0f15dab9dea.csv> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>> >>>>>>>>> >>>>>>>> >>>>>>> >>>>>> >>>>> >>>> >>>> >>> >> >> >

1 0

Enabling Spatial support?
by Thorsten Wesskallnies 17 Jun '15

17 Jun '15

Hi, I'm having some trouble with the instruction posted on the website here: https://www.monetdb.org/Documentation/Extensions/GIS I am running an OEL Server on version 6.6 and installed MonetDB using yum: Installed Packages Name : MonetDB Arch : x86_64 Version : 11.19.11 Release : 20150423.el6 ... Name : MonetDB-geom-MonetDB5 Arch : x86_64 Version : 11.19.11 Release : 20150423.el6 I also configured and installed the current version of the GEOS Library # geos-config --version >3.4.2 According to the above mentioned website I should now rebuild my DB with the '--enable-geom configure argument'. Unfortunately I'm unable to find the correct way to do it since yum installed a binary on my system and not the sourcecode. The 'MonetDB-geom' package doesn't seem to to the trick since the result of the sample Code on the Website returns an empty result. Hope somebody can point me in the right direction. MfG Thorsten Wesskallnies Avitech GmbH Principal Office: Bahnhofplatz 1 | 88045 Friedrichshafen | Germany Court Registration: Amtsgericht Ulm | HRB 728293 Geschäftsführer/Managing Director: Antonio Maria Gonzalez Gorostiza VAT No.: DE223719716 http://avitech.aero

2 1

trace disabled?
by Roberto Cornacchia 16 Jun '15

16 Jun '15

Hi there, Just to be sure: is it expected that trace is disabled / changed in Jan2015? I can see the trace by inspecting sys.tracelog() after the query, but no longer in one go as before. Roberto

1 1

Performing a quantile in R from a MonetDB.R db with 300million rows (UNCLASSIFIED)
by George, Glover E ERD-MS 15 Jun '15

15 Jun '15

Classification: UNCLASSIFIED Caveats: NONE Hi all, I'm currently trying to compare the performance of R's quantile function to that of MonetDB's. I have a table loaded with the TPC-H benchmark's data (scale factor of 50). I'm trying to return ~300 million rows of data (select l_extendedprice from lineitem), and although I have 256G of RAM, I'm getting the following: Error in paste(0)("",resp, collapse="") : Result would exceed 2^31-1 bytes The R statement is : x = dbGetQuery(conn, "select l_extendedprice from lineitem") Forgive me if this is a simplistic questions, as I am fairly new to R and don't know which avenue to take next. I took a look at bigmemory, but don't know quite how to return a result from monetdb and have it use a big.matrix. Note: the l_extendedprice is a floating (double) value. Any suggestions? Cheers. Glover George Classification: UNCLASSIFIED Caveats: NONE

2 2