- users-list - monetdb.org

Extending physical memory via memory cache
by Calum Miller 26 Mar '14

26 Mar '14

Good Morning, Is it possible to extend the amount of physical memory available to monetdb via Memcached/Hazelcast? Thanks Calum Miller Millersoft

2 1

JDBC - how to set autocommit in url as on option
by Radovan Bičiště 21 Mar '14

21 Mar '14

Hello, We are using the JDBC driver from KETTLE (http://kettle.pentaho.com) tool. I could not find an option how to tell the driver that my connections are autocommit. I tried to add option autocommit=true. The tool added the opton at the end of the JDBC URL as follows jdbc:monetdb://172.26.123.14:50000/zkdev1;autocommit=true . But it failed with error that it could not find database "zkdev1;autocommit=true". Thank you for your help. Radovan

3 2

How to execute a MAL function in a monetdb database ?
by Leon-Constantin Fopa 17 Mar '14

17 Mar '14

I'm reading the MonetDB Internal Description because I'm interested to implement a special procedure direct in the MonetDB Algebra Language (MAL) instead of SQL-frontend (I hope to gain more performance by doing so). Here is a sample MAL code provide in the MonetDB documentation function sample(nme:str, val:any_1):bit; c := 2 * 3; b := bbp.bind(nme); #find a BAT h := algebra.select(b,val,val); t := aggr.count(h); x := io.print(t); y := io.print(val); end sample; My question is how to execute such a MAL code upon one of my existing database ? Thans for any replay

2 1

Copy table between monet instances using COPY BINARY ...
by Radovan Bičiště 17 Mar '14

17 Mar '14

Hello, I read the following recipe https://www.monetdb.org/Documentation/Cookbooks/SQLrecipes/BinaryBulkLoad . Please is there a command or a binary that does the dump part "Each attachment file is produced by a program that writes the binary image of the BAT directly, ..." ? (mclient --format=raw?) We are on "MonetDB Database Server v1.7 (Feb2013-SP6)". Thank you for your help. Radovan

2 1

COPY INTO table specify columns
by Wei Zhuo 17 Mar '14

17 Mar '14

Hi, Couldn't find any info/doc if a list of column names, and order of columns, can be specified when using COPY INTO command. If not possible, is this something relatively easy to implement? Thanks!

2 1

possible to improve paging performance on the windows version of monetdb?
by Anthony Damico 17 Mar '14

17 Mar '14

hi monetdb team, could disk paging performance be improved on windows? a query run without versus with any disk paging is kinda like the hulk versus plain old bruce banner. ;) ==background== i'm working on a longer tutorial for http://asdfree.com/ about how to work with census microdata from 74 different countries (yes, 74!) with R and monetdb. i'm planning to release a blog post specifically for the brazilian microdata and then expand that code to work on the other 73 countries <https://international.ipums.org/international-action/countries>. both brazilian census bureau employees and outside users have taken a serious interest in this work, with official links from one<http://www.ibge.gov.br/home/estatistica/populacao/trabalhoerendimento/pnad2…>, two<http://www.ibge.gov.br/home/estatistica/indicadores/trabalhoerendimento/pme…>, three <http://www.ibge.gov.br/home/xml/pof_2008_2009.shtm> of their microdata pages. r and monetdb are obviously great choices for analysts who don't have the resources for expensive computers and/or software. a former employee at insee (the french stat. bureau) who currently consults on international demography projects outlined this problem for me pretty clearly: lower-resource statistical agencies in the developing world need a good, free option like (R + monetdb) for analysis of larger data sets on the (generally cheaper) hardware that they already have. international census bureau employees and other demographers are much more likely to have statistical but not computer programming skills, so as i work on these instructional documents, i'll aim to make the computing side of things as easy as possible. and, most central to this e-mail, much of this new potential audience uses windows. (on march 10th, 2014<http://cran-logs.rstudio.com/2014/2014-03-10.csv.gz>, 67% of MonetDB.R package downloads were windows users) ==problem== certain sql commands - it's hard for me to pin down which ones - tend to hang. i will run a command, look at my "process manager" and see RAM start getting gobbled up (not by mserver.exe but by some invisible process--i think disk paging?), and then wait. mserver.exe will use between 0% and 1% of the total CPU for a few minutes, then something will "wake up" and suddenly the CPU starts lighting up.. with 70%-90% usage across multiple cores. this 1% CPU usage to 75% CPU usage can jump back and forth a few times until the full command finally finishes. all of the time spent with 0% or 1% CPU usage - waiting for commands to execute - seems like an opportunity for improvement of the windows build? here's an example that took 11 minutes for a query run on 20,000,000 records.. > system.time( z <- svymean( ~v0601 , pes.design , byvar = ~v0001 , se = TRUE ) ) user system elapsed 18.47 1.13 *710.86 * ..totally awesome! but can it be improved? ==comparison between two machines== the attached .sql code has been run on two windows machines: on the windows server with 64 gb of ram-- # Serving database 'test', using 4 threads # Compiled for x86_64-pc-winnt/64bit with 64bit OIDs dynamically linked # Found 64.000 GiB available main-memory. on my personal windows laptop-- # Serving database 'test', using 8 threads # Compiled for x86_64-pc-winnt/64bit with 64bit OIDs dynamically linked # Found 7.860 GiB available main-memory. ==processing time comparison== processing time on the 64gb server: 27 tuples *(5.2s)* processing time on my 8gb laptop: 27 tuples *(7m 46s)* *so on a windows machine, if disk paging can be avoided, the processing time is 90x faster for this attached sql code?* ==how to load this reproducible example== ** data file with four columns ** https://drive.google.com/file/d/0B3LuuP081KDYNERWQWx5MEpZU2M/edit?usp=shari… ** commands to load the data ** CREATE TABLE "sys"."d10" ( "idkey" INTEGER, "v0601" VARCHAR(255), "pes_wgt" DOUBLE, "v0001" VARCHAR(255) ); COPY 20635472 RECORDS INTO d10 FROM 's:/temp/example.txt'; [[see .sql attachment for post-import commands]] ==general diagnostics== # MonetDB 5 server v11.17.11 "Jan2014-SP1" # Serving database 'test', using 4 threads # Compiled for x86_64-pc-winnt/64bit with 64bit OIDs dynamically linked # Found 64.000 GiB available main-memory. # Copyright (c) 1993-July 2008 CWI. # Copyright (c) August 2008-2014 MonetDB B.V., all rights reserved # Visit http://www.monetdb.org/ for further information # Listening for connection requests on mapi:monetdb://127.0.0.1:50000/ # MonetDB/JAQL module loaded # MonetDB/SQL module loaded > ==conclusion question== the windows version of monetdb is intensely slow at rather unexpected times. can and will anything be done about this in the near future? thanks for monetdb!

1 1

MonetDB Slow Query with Index
by Summer Xia 15 Mar '14

15 Mar '14

Hi all, I created a table with 10GB (60M records), added unique index (hidden_id) manually after data insertion. I had the simplest query but it took one minutes to complete. select hidden_id from netflow where hidden_id = 350000; And also the query took tens of minutes "select * from netflow order by hidden_id limit 12500 offset 212500;". It really confuses me. I post the analysis of the first below. Any clue why it's so slow? trace select hidden_id from netflow where hidden_id = 350000; +----------+------------------------------------------------------------------+ | ticks | stmt | +==========+==================================================================+ | 3 | X_3 := sql.mvc(); | | 15 | X_7=<tmp_2510>[69396995] := sql.bind(X_3=0,"sys","netflow","hidd | : : en_id",0); : | 227 | X_4:bat[:oid,:oid] =<tmp_13332>[69396995] := sql.tid(X_3=0,"sys" | : : ,"netflow"); : | 72978741 | X_36=<tmp_4053>[1] := algebra.subselect(X_7=<tmp_2510>[69396995] | : : ,X_4=<tmp_13332>:bat[:oid,:oid][69396995],A0=350000:lng,A0=35000 : : : 0:lng,true,true,false); : | 17 | (X_10=<tmp_2175>[0],r1_10=<tmp_3416>[0]) := sql.bind(X_3=0,"sys" | : : ,"netflow","hidden_id",2); : | 14 | X_37=<tmp_13337>[0] := algebra.subselect(r1_10=<tmp_3416>[0],A0= | : : 350000:lng,A0=350000:lng,true,true,false); : | 6 | X_13=<tmp_3416>[0] := sql.bind(X_3=0,"sys","netflow","hidden_id" | : : ,1); : | 15 | X_38=<tmp_11053>[0] := algebra.subselect(X_13=<tmp_3416>[0],X_4= | : : <tmp_13332>:bat[:oid,:oid][69396995],A0=350000:lng,A0=350000:lng : : : ,true,true,false); : | 4 | X_15=<tmp_4053>[1] := sql.subdelta(X_36=<tmp_4053>[1],X_4=<tmp_1 | : : 3332>:bat[:oid,:oid][69396995],X_10=<tmp_2175>[0],X_37=<tmp_1333 : : : 7>[0],X_38=<tmp_11053>[0]); : | 20 | X_17=<tmp_11053>[1] := sql.projectdelta(X_15=<tmp_4053>[1],X_7=< | : : tmp_2510>[69396995],X_10=<tmp_2175>[0],r1_10=<tmp_3416>[0],X_13= : : : <tmp_3416>[0]); : | 6 | X_18 := sql.resultSet(1,1,X_17=<tmp_11053>[1]); | | 7 | sql.rsColumn(X_18=1,"sys.netflow","hidden_id","bigint",64,0,X_17 | : : =<tmp_11053>[1]); : | 2 | X_23 := io.stdout(); | | 25 | sql.exportResult(X_23=="104d2":streams,X_18=1); | | 1 | end s1_3; | | 73011629 | X_5:void := user.s1_3(350000:lng); | +----------+------------------------------------------------------------------+ This is the table being created. CREATE TABLE "netflow" ( "time_seconds" double DEFAULT NULL, "parsed_date" timestamp DEFAULT NULL, "date_time_str" varchar(45) DEFAULT NULL, "ip_layer_protocol" bigint DEFAULT NULL, "ip_layer_protocol_code" varchar(45) DEFAULT NULL, "first_seen_src_ip" varchar(45) DEFAULT NULL, "first_seen_dest_ip" varchar(45) DEFAULT NULL, "first_seen_src_port" bigint DEFAULT NULL, "first_seen_dest_port" bigint DEFAULT NULL, "more_fragments" varchar(45) DEFAULT NULL, "cont_fragments" varchar(45) DEFAULT NULL, "duration_seconds" bigint DEFAULT NULL, "first_seen_src_payload_bytes" bigint DEFAULT NULL, "first_seen_dest_payload_bytes" bigint DEFAULT NULL, "first_seen_src_total_bytes" bigint DEFAULT NULL, "first_seen_dest_total_bytes" bigint DEFAULT NULL, "first_seen_src_packet_count" bigint DEFAULT NULL, "first_seen_dest_packet_count" bigint DEFAULT NULL, "record_force_out" varchar(45) DEFAULT NULL ); Best regards, Summer

2 1

python 3.2 support
by Simon 13 Mar '14

13 Mar '14

Hi thanks for the great project. I just wonder when there would be python 3.2 support. Is there any way that we could use monetdb in python 3.2 before the official package? Thanks again. Simon

4 5

Announcement: New Jan2014-SP1 Bugfix release of MonetDB suite
by Sjoerd Mullender 10 Mar '14

10 Mar '14

The MonetDB team at CWI/MonetDB BV is pleased to announce the Jan2014-SP1 bugfix release of the MonetDB suite of programs. More information about MonetDB can be found on our website at <http://www.monetdb.org/>. For details on this release, please see the release notes at <http://www.monetdb.org/Downloads/ReleaseNotes>. As usual, the download location is <http://dev.monetdb.org/downloads/>. Jan 2014-SP1 bugfix release Build Environment * Configure now enables the SQL front end by default. Merovingian * monetdb destroy -f now also works on running databases Bug Fixes * 3442: COPY INTO ... LOCKED reports incorrect count * 3443: DROP INDEX crashes server with BATsubselect: invalid argument: b must have a dense head * 3444: AND after ON () of LEFT OUTER JOIN with certain expressions will cause crash * 3452: ODBC driver build fails on Mac OS X due to a conflicting types for the SQLColAttribute with the unixODBC library

1 0

Fwd: [asdfree by anthony damico] New comment on analyze the program for international student asse....
by Anthony Damico 04 Mar '14

04 Mar '14

if there's any chance you might release SP1 soon, i'd appreciate it :) this user is hitting (resolved) bug 3443<http://bugs.monetdb.org/show_bug.cgi?id=3443>. thanks! ---------- Forwarded message ---------- From: lightest <noreply-comment(a)blogger.com> Date: Sun, Feb 23, 2014 at 7:53 PM Subject: [asdfree by anthony damico] New comment on analyze the program for international student asse.... To: ajdamico(a)gmail.com lightest <http://www.blogger.com/profile/12197082513968468249> has left a new comment on your post "analyze the program for international student asse...<http://www.asdfree.com/2013/12/analyze-program-for-international.html>": Hi Anthony, I've tried to run to codes, and when I run lines such as those of MICombine() functions, sometimes there is an error like the following: In R: Error in .mapiRequest(conn, paste0("s", statement, ";"), async = async) : error writing to socket (10054) Error in .mapiRequest(conn, paste0("s", statement, ";"), async = async) : error writing to socket (10054) Error in .mapiRequest(conn, paste0("s", statement, ";"), async = async) : error writing to socket (10054) > ======================================== In the console: # MonetDB 5 server v11.17.9 "Jan2014" # Serving database 'pisa', using 4 threads # Compiled for x86_64-pc-winnt/64bit with 64bit OIDs dynamically linked # Found 7.913 GiB available main-memory. # Copyright (c) 1993-July 2008 CWI. # Copyright (c) August 2008-2014 MonetDB B.V., all rights reserved # Visit http://www.monetdb.org/ for further information # Listening for connection requests on mapi:monetdb://127.0.0.1:50007/ # MonetDB/JAQL module loaded # MonetDB/SQL module loaded >!FATAL: 40000!COMMIT: transation commit failed (perhaps your disk is full?) exi ting (kernel error: !ERROR: BATsubselect: invalid argument: b must have a dense head. ) Press any key to continue . . . ======================================= Publish<http://www.blogger.com/comment-moderate-confirm.do?blogID=69830612920483473…> Delete<http://www.blogger.com/comment-moderate-confirm.do?blogID=69830612920483473…> Mark as spam<http://www.blogger.com/comment-moderate-confirm.do?blogID=69830612920483473…> Moderate<http://www.blogger.com/blogger.g?blogID=6983061292048347330#pendingcomments>comments for this blog. Posted by lightest to asdfree by anthony damico <http://www.asdfree.com/>at February 23, 2014 at 7:53 PM

2 3