users-list April 2008

users-list@monetdb.org

21 participants
31 discussions

by Matt Magoffin

Hello, I am interested in evaluating MonetDB/XQuery. I have some XML collections for a project I work on that are of 1-2 million individual XML documents for which I use XQuery to access for OLAP style reporting operations. This is a Java-based project, and while working on this project I developed a thin XQuery-centric API so that I could evaluate/use many different XQuery implementations, which my company released as an open source project called Xaj (http://sourceforge.net/projects/xaj). Currently I am using a Berkeley DB XML implementation, but am interested in exploring other options. I'd like to implement a MonetDB version of this API now, but I am having some difficulty finding out how to effectively access MonetDB through Java. So far I was able to use the MonetDB JDBC driver to construct XQuery statements and get XML results, and also I was able to construct MonetDB pf:add-doc() statements to add documents, but I am wondering if this is the best approach (and if so, how to add documents to a remote MonetDB server using the pf:add-doc() mechanism). The API I am trying to implement you can view here, to give you an idea what I'm trying to do: http://xaj.svn.sourceforge.net/viewvc/xaj/xaj/src/net/sf/xaj/XmlDb.java?rev… It was somewhat modeled after the defunct XML:DB API, but focused just on add/store/XQuery operations. Any help/advise would be greatly appreciated! -- m@

13 years, 11 months

[MonetDB-users] Performance issues with queries in SQL

by Nils Grimsmo

I have tables with around 12 million rows and large primary keys. I find the performance for various queries to be rather strange. * Why is Q2 (see below) so much slower than Q1? It is actually faster for me to use Q1 and do the rest of the matching myself. * Why is Q3 so immensely much faster than Q1? I understand that the first part of the primary key is selective in Q3. Do I understand right if Q1 scans until "type" matches, then scans until "name" matches, then scans until "value" matches? No indexes are put on disk, right? Then the matching for Q3 should be scan until "value" matches, which shouldn't be more than at most three times faster, right (if three columns are read instead of one for Q1)? I do Q1 twice to look for cash effects, but none are seen... Could some columns from the table "value" be cached and kept, without giving room for those from "type_name_value"? * Q4 has the same performance as Q1, as expected. * Why isn't Q5 faster than Q6? 82% of the rows have type=3. Fewer should be scanned to find that "!!foo" doesn't exist than for "~~foo", as "!" has a low ASCII value, while "~" is high. (See COUNT queries at the bottom.) Klem fra Nils Q1: SELECT * FROM type_name_value WHERE type = 3 AND name = 'author' AND value = 'Jurgen Annevelink' Q2: SELECT * FROM type_name_value WHERE type = 3 AND name = 'author' AND value = 'Jurgen Annevelink' AND doc = 1 AND nenc = '1.1.1.1' AND pos = 4 Q3: SELECT * FROM value WHERE value = 'Jurgen Annevelink' AND type = 3 AND name = 'author' Q4: SELECT * FROM type_name_value WHERE type = 3 AND name = 'author' AND value = '__bar' Q5: SELECT * FROM type_name_value WHERE type = 3 AND name = '!!foo' AND value = '__bar' Q6: SELECT * FROM type_name_value WHERE type = 3 AND name = '~~foo' AND value = '__bar' Q7: SELECT * FROM type_name_value WHERE type = 3000 AND name = '__foo' AND value = '__bar' See runs below. Number of rows SELECT COUNT(type) FROM type_name_value % nilsgri.type_name_value # table_name % count_no_nil_type # name % int # type % 8 # length [ 12165109 ] Query time: 0.066192 SELECT COUNT(type) FROM value % nilsgri.value # table_name % count_no_nil_type # name % int # type % 8 # length [ 12165109 ] Query time: 0.058111 The PRIMARY KEY of type_name_value is (type,name,value,doc,nenc,pos) The PRIMARY KEY of value is (value,doc,nenc,pos) Q1: Search 3 first, match 3 first. SELECT * FROM type_name_value WHERE type = 3 AND name = 'author' AND value = 'Jurgen Annevelink' % nilsgri.type_name_value, nilsgri.type_name_value, nilsgri.type_name_value, nilsgri.type_name_value, nilsgri.type_name_value, nilsgri.type_name_value, nilsgri.type_name_value, nilsgri.type_name_value # table_name % type, name, value, doc, nenc, pos, pathid, complete # name % int, varchar, clob, int, varchar, int, int, boolean # type % 1, 6, 17, 1, 12, 8, 3, 5 # length [ 3, "author", "Jurgen Annevelink", 1, "1.1.1.1", 4, 5, true ] [ 3, "author", "Jurgen Annevelink", 1, "1.23579.2.1", 724888, 56, true ] [ 3, "author", "Jurgen Annevelink", 1, "1.2487.3.1", 81996, 5, true ] [ 3, "author", "Jurgen Annevelink", 1, "1.42.2.1", 1071, 5, true ] [ 3, "author", "Jurgen Annevelink", 1, "1.533986.2.1", 17294725, 56, true ] [ 3, "author", "Jurgen Annevelink", 1, "1.576139.3.1", 18745432, 56, true ] [ 3, "author", "Jurgen Annevelink", 1, "1.944085.1.1", 30104242, 107, true ] [ 3, "author", "Jurgen Annevelink", 1, "1.99173.2.1", 3191470, 56, true ] Query time: 1.643438 Q1: Repeat to check for disk cache effects. SELECT * FROM type_name_value WHERE type = 3 AND name = 'author' AND value = 'Jurgen Annevelink' % nilsgri.type_name_value, nilsgri.type_name_value, nilsgri.type_name_value, nilsgri.type_name_value, nilsgri.type_name_value, nilsgri.type_name_value, nilsgri.type_name_value, nilsgri.type_name_value # table_name % type, name, value, doc, nenc, pos, pathid, complete # name % int, varchar, clob, int, varchar, int, int, boolean # type % 1, 6, 17, 1, 12, 8, 3, 5 # length [ 3, "author", "Jurgen Annevelink", 1, "1.1.1.1", 4, 5, true ] [ 3, "author", "Jurgen Annevelink", 1, "1.23579.2.1", 724888, 56, true ] [ 3, "author", "Jurgen Annevelink", 1, "1.2487.3.1", 81996, 5, true ] [ 3, "author", "Jurgen Annevelink", 1, "1.42.2.1", 1071, 5, true ] [ 3, "author", "Jurgen Annevelink", 1, "1.533986.2.1", 17294725, 56, true ] [ 3, "author", "Jurgen Annevelink", 1, "1.576139.3.1", 18745432, 56, true ] [ 3, "author", "Jurgen Annevelink", 1, "1.944085.1.1", 30104242, 107, true ] [ 3, "author", "Jurgen Annevelink", 1, "1.99173.2.1", 3191470, 56, true ] Query time: 1.615900 Q2: Search entire, match entire. SELECT * FROM type_name_value WHERE type = 3 AND name = 'author' AND value = 'Jurgen Annevelink' AND doc = 1 AND nenc = '1.1.1.1' AND pos = 4 % nilsgri.type_name_value, nilsgri.type_name_value, nilsgri.type_name_value, nilsgri.type_name_value, nilsgri.type_name_value, nilsgri.type_name_value, nilsgri.type_name_value, nilsgri.type_name_value # table_name % type, name, value, doc, nenc, pos, pathid, complete # name % int, varchar, clob, int, varchar, int, int, boolean # type % 1, 6, 17, 1, 7, 1, 1, 5 # length [ 3, "author", "Jurgen Annevelink", 1, "1.1.1.1", 4, 5, true ] Query time: 11.073827 Q3: Use table value instead. SELECT * FROM value WHERE value = 'Jurgen Annevelink' AND type = 3 AND name = 'author' % nilsgri.value, nilsgri.value, nilsgri.value, nilsgri.value, nilsgri.value, nilsgri.value, nilsgri.value, nilsgri.value # table_name % value, doc, nenc, pos, type, name, pathid, complete # name % clob, int, varchar, int, int, varchar, int, boolean # type % 17, 1, 12, 8, 1, 6, 3, 5 # length [ "Jurgen Annevelink", 1, "1.1.1.1", 4, 3, "author", 5, true ] [ "Jurgen Annevelink", 1, "1.23579.2.1", 724888, 3, "author", 56, true ] [ "Jurgen Annevelink", 1, "1.2487.3.1", 81996, 3, "author", 5, true ] [ "Jurgen Annevelink", 1, "1.42.2.1", 1071, 3, "author", 5, true ] [ "Jurgen Annevelink", 1, "1.533986.2.1", 17294725, 3, "author", 56, true ] [ "Jurgen Annevelink", 1, "1.576139.3.1", 18745432, 3, "author", 56, true ] [ "Jurgen Annevelink", 1, "1.944085.1.1", 30104242, 3, "author", 107, true ] [ "Jurgen Annevelink", 1, "1.99173.2.1", 3191470, 3, "author", 56, true ] Query time: 0.000477 Q4: Search 3 first, match 2 first. SELECT * FROM type_name_value WHERE type = 3 AND name = 'author' AND value = '__bar' % nilsgri.type_name_value, nilsgri.type_name_value, nilsgri.type_name_value, nilsgri.type_name_value, nilsgri.type_name_value, nilsgri.type_name_value, nilsgri.type_name_value, nilsgri.type_name_value # table_name % type, name, value, doc, nenc, pos, pathid, complete # name % int, varchar, clob, int, varchar, int, int, boolean # type % 1, 0, 0, 1, 0, 1, 1, 5 # length Query time: 1.617259 Q5: Search 3 first, match 1 first. '!' early in alphabet. SELECT * FROM type_name_value WHERE type = 3 AND name = '!!foo' AND value = '__bar' % nilsgri.type_name_value, nilsgri.type_name_value, nilsgri.type_name_value, nilsgri.type_name_value, nilsgri.type_name_value, nilsgri.type_name_value, nilsgri.type_name_value, nilsgri.type_name_value # table_name % type, name, value, doc, nenc, pos, pathid, complete # name % int, varchar, clob, int, varchar, int, int, boolean # type % 1, 0, 0, 1, 0, 1, 1, 5 # length Query time: 0.530727 Q6: Search 3 first, match 1 first. '~' late in alphabet. SELECT * FROM type_name_value WHERE type = 3 AND name = '~~foo' AND value = '__bar' % nilsgri.type_name_value, nilsgri.type_name_value, nilsgri.type_name_value, nilsgri.type_name_value, nilsgri.type_name_value, nilsgri.type_name_value, nilsgri.type_name_value, nilsgri.type_name_value # table_name % type, name, value, doc, nenc, pos, pathid, complete # name % int, varchar, clob, int, varchar, int, int, boolean # type % 1, 0, 0, 1, 0, 1, 1, 5 # length Query time: 0.534592 Q7: Search 3 first, match 0 first. SELECT * FROM type_name_value WHERE type = 3000 AND name = '__foo' AND value = '__bar' % nilsgri.type_name_value, nilsgri.type_name_value, nilsgri.type_name_value, nilsgri.type_name_value, nilsgri.type_name_value, nilsgri.type_name_value, nilsgri.type_name_value, nilsgri.type_name_value # table_name % type, name, value, doc, nenc, pos, pathid, complete # name % int, varchar, clob, int, varchar, int, int, boolean # type % 1, 0, 0, 1, 0, 1, 1, 5 # length Query time: 0.000407 Number of rows with type and name mathcing. SELECT COUNT(type) FROM type_name_value WHERE type = 3 AND name = 'author' % nilsgri.type_name_value # table_name % count_no_nil_type # name % int # type % 7 # length [ 2730959 ] Query time: 0.683863 Number of rows with type mathcing. SELECT COUNT(type) FROM type_name_value WHERE type = 3 % nilsgri.type_name_value # table_name % count_no_nil_type # name % int # type % 8 # length [ 10020569 ] Query time: 0.032573 SELECT COUNT(type) FROM type_name_value WHERE type = 3 AND name < '!!foo' % nilsgri.type_name_value # table_name % count_no_nil_type # name % int # type % 1 # length [ 0 ] Query time: 0.161186 SELECT COUNT(type) FROM type_name_value WHERE type = 3 AND name < '~~foo' % nilsgri.type_name_value # table_name % count_no_nil_type # name % int # type % 8 # length [ 10020569 ] Query time: 0.341732 Number of rows with type mathcing. real 0m18.372s user 0m0.005s sys 0m0.002s

14 years, 2 months

[MonetDB-users] How to set the memory limit for monetDB.

by Kang Zhang

I am using the latest version of monetDB from the web-site(the super source tarball). My machine is a X86-64 machine with a red-hat installed on it. When I insert data into monetDB, it always requires memory from the system. Then it will be crashed with running out of memory. Is there a way to set the limit for memory used by monetDB ? I've read the docs on the web site and can't get what I need from that. Thank you. -- ------------------------------------------------------------ Kang Zhang Computer Science Dept. Shanghai Jiao Tong University

14 years, 2 months

[MonetDB-users] Problems with parameters to mapi_param*()

by Nils Grimsmo

I am having some problems with the functions mapi_param*(). 1. It seems mapi_param() really wants a char* cast to a char**. (See the inner workings of mapi_param() vs. mapi_param_string().) 2. If mapi_param(), mapi_param_string() and mapi_param_type() get a char*, the value used in the query does not change as the char* is updated. Using mapi_bind_var( , , MAPI_INT, int*) and mapi_bind_var( , , MAPI_VARCHAR, char**) works as expected, while mapi_param_type( , , MAPI_INT, MAPI_INT, int*) and mapi_param_type( , , MAPI_VARCHAR, MAPI_VARCHAR, char**) do not. Is there something wrong with interface, or am I just using it wrong? Here is how I have tried using the functions: #include <cassert> #include <cstdio> #include <cstdlib> #include <cstring> #include <Mapi.h> using namespace std; Mapi dbh; void check(MapiHdl hdl, int ret = MOK, bool fatal = true) { if (ret != MOK || hdl == NULL || mapi_error(dbh)) { if (hdl) { mapi_explain_result(hdl,stderr); } else { mapi_explain(dbh,stderr); } if (fatal) assert(0);//exit(-1); } } void paramInt(MapiHdl hdl, int fldnr, int* val) { int ret = mapi_param_type(hdl, fldnr, MAPI_INT, MAPI_INT, val); check(hdl, ret); } void paramStr(MapiHdl hdl, int fldnr, char** val) { // int ret = mapi_param(hdl, fldnr, val); /* Case 1: * QUERY = INSERT INTO t VALUES(1, p`) * ERROR = !SELECT: identifier 'p`' unknown * Cast to (void*) in call to mapi_param_type() in mapi_param(). */ int ret = mapi_param(hdl, fldnr, (char**)*val); /* Case 2: * Works, sort of. * Seems it wants a (char*) really. * But 'foo' is inserted the second time instead of bar. */ // int ret = mapi_param_string(hdl, fldnr, MAPI_VARCHAR, *val, NULL); /* 'foo' twice. */ // int ret = mapi_param_string(hdl, fldnr, MAPI_VARCHAR, (char*)val, NULL); /* As Case 1. */ // int ret = mapi_param_type(hdl, fldnr, MAPI_VARCHAR, MAPI_VARCHAR, val); /* As Case 1. */ // int ret = mapi_param_type(hdl, fldnr, MAPI_VARCHAR, MAPI_VARCHAR, *val); /* 'foo' twice. */ check(hdl, ret); } int main() { dbh = mapi_connect("localhost", 50000, "nilsgri", "nilsgri", "sql", "proto"); MapiHdl hdl = mapi_query(dbh, "DROP TABLE t"); check(hdl, MOK, false); mapi_close_handle(hdl); hdl = mapi_query(dbh, "CREATE TABLE t (a INT, b VARCHAR(255))"); check(hdl); mapi_close_handle(hdl); hdl = mapi_prepare(dbh, "INSERT INTO t VALUES(?, ?)"); check(hdl); int a = 1; paramInt(hdl, 0, &a); char* b1 = strdup("'foo'"); char* b = b1; paramStr(hdl, 1, &b); int ret = mapi_execute(hdl); check(hdl, ret); a = 2; char* b2 = strdup("'bar'"); b = b2; // paramStr(hdl, 1, &b); // I would like not having to do this. ret = mapi_execute(hdl); check(hdl, ret); mapi_close_handle(hdl); hdl = mapi_quick_query(dbh, "SELECT * FROM t", stderr); check(hdl, ret); mapi_close_handle(hdl); mapi_disconnect(dbh); mapi_destroy(dbh); free(b1); free(b2); }

14 years, 3 months

[MonetDB-users] Is it a bug? (insert into ... select command)

by Iftikharn

Hi I need to insert some null values into a table via a procedure. I created the table and did insert two rows. One with (insert into) command and second with (insert into ..select) command. Both commands work well and I could see two rows added in result. When I used the same (insert into.. select) command in a procedure. It behaves in a strange fashion. It shows that a row has been affected, but does not show the newly added row. If the rows are counted, it gives the correct count which is (3) in this case but still I could not see the last row which is added via a procedure. ''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''' create sequence "ff" as integer start with 1; CREATE TABLE t1(timeid INT DEFAULT NEXT VALUE FOR "ff" PRIMARY KEY, a INT, b INT, c INT, d INT, e INT, f VARCHAR(20), g VARCHAR(20), h INT, i VARCHAR(10), j VARCHAR(10), k VARCHAR(10), l INT, m INT, n VARCHAR(10), o VARCHAR(10), p VARCHAR(10), q VARCHAR(15), r INT ); ''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''' And then insert a row ''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''' insert into t1(a, b, c, d, e, f, g, h, i, j, k, l, m, n, o, p, q, r) values(1, 1, 1, 1, 1, 'first row', 'first row',1, 'first row', 'first row', 'first row',1,1, 'first row', 'first row', 'first row', 'first row', 1); select * from t1;sql>sql>sql>sql>more>more>more>more>more>more>more>more>more>more>more>more>more>more>more>more>more>more>more>sql> sql>Rows affected 1 sql> +------+--+--+--+--+--+-----+-----+--+-----+------+------+--+--+------+------+------+------+--+ | time |a |b |c |d |e |f |g |h |i |j |k |l |m |n |o |p |q |r | : id | | | | | | | | | | | | | | | | | | | +======+==+==+==+==+==+=====+=====+==+=====+======+======+==+==+======+======+======+======+==+ | 1 |1 |1 |1 |1 |1 |firs |firs |1 |firs |first |first |1 |1 |first |first |first |first |1 | : | | | | | |t |t | |t |row |row | | |row |row |row |row | | : | | | | | |row |row | |row | | | | | | | | | | +------+--+--+--+--+--+-----+-----+--+-----+------+------+--+--+------+------+------+------+--+ '''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''' Every thing is fine so far '''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''' sql>insert into t1(a, b, c, d, e, f, g, h) select a, b, c, d, e, f, g, h from t1 where timeid = 1; Rows affected 1 sql>select * from t1; +------+--+--+--+--+--+-----+-----+--+-----+------+------+--+--+------+------+------+------+--+ | time |a |b |c |d |e |f |g |h |i |j |k |l |m |n |o |p |q |r | : id | | | | | | | | | | | | | | | | | | | +======+==+==+==+==+==+=====+=====+==+=====+======+======+==+==+======+======+======+======+==+ | 1 |1 |1 |1 |1 |1 |firs |firs |1 |firs |first |first |1 |1 |first |first |first |first |1 | : | | | | | |t |t | |t |row |row | | |row |row |row |row | | : | | | | | |row |row | |row | | | | | | | | | | | 2 |1 |1 |1 |1 |1 |firs |firs |1 |null |null |null |n |n |null |null |null |null |n | : | | | | | |t |t | | | | |u |u | | | | |u | : | | | | | |row |row | | | | |l |l | | | | |l | : | | | | | | | | | | | |l |l | | | | |l | +------+--+--+--+--+--+-----+-----+--+-----+------+------+--+--+------+------+------+------+--+ ''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''' inserted another row and it works fine '''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''' sql>create procedure tt() begin insert into t1(a, b, c, d, e, f, g, h) select a, b, c, d, e, f, g, h from t1 where timeid = 1; end; call tt(); select * from t1; more>more>more> sql>Rows affected 1 sql>+------+--+--+--+--+--+-----+-----+--+-----+------+------+--+--+------+------+------+------+--+ | time |a |b |c |d |e |f |g |h |i |j |k |l |m |n |o |p |q |r | : id | | | | | | | | | | | | | | | | | | | +======+==+==+==+==+==+=====+=====+==+=====+======+======+==+==+======+======+======+======+==+ | 1 |1 |1 |1 |1 |1 |firs |firs |1 |firs |first |first |1 |1 |first |first |first |first |1 | : | | | | | |t |t | |t |row |row | | |row |row |row |row | | : | | | | | |row |row | |row | | | | | | | | | | | 2 |1 |1 |1 |1 |1 |firs |firs |1 |null |null |null |n |n |null |null |null |null |n | : | | | | | |t |t | | | | |u |u | | | | |u | : | | | | | |row |row | | | | |l |l | | | | |l | : | | | | | | | | | | | |l |l | | | | |l | +------+--+--+--+--+--+-----+-----+--+-----+------+------+--+--+------+------+------+------+--+ ''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''' The problem starts now: when I try to insert the row via a procedure it shows that row effected 1 but does not show the row. ''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''' sql>select count(*) from t1; +--------+ | count_ | +========+ | 3 | +--------+ sql>select timeid from t1; +--------+ | timeid | +========+ | 1 | | 2 | | 3 | +--------+ sql>select * from t1; +------+--+--+--+--+--+-----+-----+--+-----+------+------+--+--+------+------+------+------+--+ | time |a |b |c |d |e |f |g |h |i |j |k |l |m |n |o |p |q |r | : id | | | | | | | | | | | | | | | | | | | +======+==+==+==+==+==+=====+=====+==+=====+======+======+==+==+======+======+======+======+==+ | 1 |1 |1 |1 |1 |1 |firs |firs |1 |firs |first |first |1 |1 |first |first |first |first |1 | : | | | | | |t |t | |t |row |row | | |row |row |row |row | | : | | | | | |row |row | |row | | | | | | | | | | | 2 |1 |1 |1 |1 |1 |firs |firs |1 |null |null |null |n |n |null |null |null |null |n | : | | | | | |t |t | | | | |u |u | | | | |u | : | | | | | |row |row | | | | |l |l | | | | |l | : | | | | | | | | | | | |l |l | | | | |l | +------+--+--+--+--+--+-----+-----+--+-----+------+------+--+--+------+------+------+------+--+ sql> ''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''' I could not see the 3rd row *********************************** Regards, Nadeem -- View this message in context: http://www.nabble.com/Is-it-a-bug--%28insert-into-...-select-command%29-tp1… Sent from the monetdb-users mailing list archive at Nabble.com.

14 years, 3 months

[MonetDB-users] MServer crashing: Random Segmentation fault

by jz＠uva

Hi all, I am using MD4 with XQuery + PF/Tijah module. I have a problem that my server sometimes/frequently crashes. I could not detect a trigger, as it occurs randomly and at any time. It gives me this error: 26747 Segmentation fault Mserver --dbinit="module(pathfinder);" but it gives me any number before it. I am not sure where I can locate the MD error log (if it exists), and how it is triggered. Help? Thanks, junte

14 years, 3 months

[MonetDB-users] MALException:opt.remap:no implementation present

by McKennirey.Matthew

Our application includes multiple services that independently connect to the database. When each service connects it reads the database metadata to discover the column names, many services may start at about the same time. We are getting the following error: java.sql.SQLException: TypeException:user.s7172_2[262]:'io.stdout' undefined in: _416:any MALException:opt.remap:no implementation present at nl.cwi.monetdb.jdbc.MonetConnection$ResponseList.executeQuery(MonetConnection.java:2058) at nl.cwi.monetdb.jdbc.MonetConnection$ResponseList.processQuery(MonetConnection.java:1808) at nl.cwi.monetdb.jdbc.MonetStatement.internalExecute(MonetStatement.java:371) at nl.cwi.monetdb.jdbc.MonetStatement.execute(MonetStatement.java:340) at nl.cwi.monetdb.jdbc.MonetStatement.executeQuery(MonetStatement.java:404) at nl.cwi.monetdb.jdbc.MonetDatabaseMetaData.getTypeInfo(MonetDatabaseMetaData.java:2454) # check status of database /monetdb status name state uptime crashes BUSINESS running 1d 1h 40m 38s 3 2008-04-22 11:23:59 # try a query using mclient running on the same machine as the database ./mclient -dBUSINESS -umonetdb -Pmonetdb -lsql sql>select * from tables; !TypeException:user.s5301_3[130]:'io.stdout' undefined in: _193:any sql>select * from schemas; !TypeException:user.s5302_3[40]:'io.stdout' undefined in: _63:any sql>select * from columns; !TypeException:user.s5303_3[146]:'io.stdout' undefined in: _210:any sql>\q; When we stop and restart the merovingian demon, the problem goes away, but may reoccur the next time we try and start multiple services in short succession. Environment MonetDB MSG merovingian[12454]: starting database 'BUSINESS', up min/avg/max: 121/43165/92925, crash average: 0.00 0.30 0.10 (7-4=3) MSG BUSINESS[12459]: executing '/monetdb/bin/mserver5' for database 'BUSINESS' MSG BUSINESS[12459]: # MonetDB server v5.5.0, based on kernel v1.23.0 MSG BUSINESS[12459]: # Serving database 'BUSINESS' MSG BUSINESS[12459]: # Compiled for x86_64-unknown-linux-gnu/64bit with 64bit OIDs dynamically linked JDBC Driver 1.7 (Canephora_p2 20080418 based on MCL v1.3)

14 years, 3 months

Re: [MonetDB-users] very slow query execution

by ewilde

hi, first: about that bug, i'll defiantly file it. second: when i migrated this table from mssql to monetdb5 I intentionally inserted it in that order (recid,ft,val). i used COPY INTO while supplying a file that i prepared in advanced with such a sorting. when i first decided to test monetdb5 i did it because i understood it should perform better than row-oriented DBMSs (like mssql), especially in CRM-like applications. so... i now understand that monetsdb5 is actually being expected to perform BAD in that specific situation i described before (in my previous message). from my experience in building CRM applications, that kind of "left join" according to unique number (recid) is very common in CRM applications. am i missing something here?? Thanks again, ewilde.

14 years, 3 months

[MonetDB-users] mserver5 process locked unkillable

by mobigital1

I was testing some queries, and for example this query has made the mserver5 process 100% CPU, which i could not kill and machine could not be rebooted. running Fedora 8 64 bit. query like this: select col1, col2, sum(col3), sum(col4), sum(col5) from thetable group by col1, col2; there are about 160 columns in the table. col1 and col2 are "smallint" col3, col4, col5 are "real" number of distinct values: in col1 is about 120 in col2 about 80 total number of rows in table is about 400M I tested the query 2 times, simpler queries work ok, like this: select col1, count(*) from thetable group by col1; this query has crashed mserver5 before, not sure if this already fixed: select col1, count(1) from thetable group col1; -- View this message in context: http://www.nabble.com/mserver5-process-locked-unkillable-tp16789362p1678936… Sent from the monetdb-users mailing list archive at Nabble.com.

14 years, 3 months

[MonetDB-users] very slow query execution

by ewilde

hello, this is how my table - table1(int,int,varchar) - looks like: recid | ft | val --------------------------------------- 1001 | 301 | a 1001 | 302 | b 1001 | 303 | c 1002 | 301 | d 1002 | 302 | e 1003 | 303 | f 1003 | 301 | g 1003 | 302 | h 1004 | 303 | i in reality this table has a few millions of records. for the following query: select t1.val as val1 , t2.val as val2 from table1 t1 left join table1 t2 on t1.recid=t2.recid and t2.ft=303 where t1.recid=1002 and t1.ft=301 i get the expected result: val1 | val2 --------------------- d | f the execution time of this query is very slow: ~1.5 seconds - in comparison to MSSql:0.016 seconds. the kind of a query that i really need in my production application doesn't have only one such a "left join...", but may have even 10. with 5 such "left join..."s the query execution time is ~25 seconds, which is so far slower than MSSql: 0.156 more info: 1. i run the queries from SQuirreL SQL Client. the MonetDB5 is installed\run on the same machine. the MSSql server was running on another machine in the same LAN. 2. the primary key of table1 is : recid,ft,val 3. at the beginning i had another problem, that i have got an empty result when running a simple query like this: select * from table1 t1 where t1.recid=1002 and t1.ft=301 i'm not sure about this, but i guess that this problem had been solved by creating a view such as: create view v1 as select recid,ft,val from table1 i have no clue whether these two problems related to each other, but i mentioned it since it might give you a hint for solving the other problem. thanks in advanced for any help, ewilde.

14 years, 3 months

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

users-list April 2008