- users-list - monetdb.org

FW: Cancel query
by Zingman, Jonathan 13 Jun '19

13 Jun '19

Yes, sys.queue(), sorry. A couple of clarifications. The cookbook shows mclient, but we're using the API and we assume you can do the same thing from the API as well. Also, we use autogenerated queries that can be hundreds if not thousands of characters long. Matching in a case independent way would be very resource intensive. It's fairly clear from the documentation that MonetDB doesn't stop immediately in the middle of whatever it is doing, so that's understood. But cancelling queries has to be part of a robust database system. My model of how this might work is that the handle that’s returned when we issue a query could be used to associate with some data returned by sys.queue(), preferably tag or start time since those should be unique and fairly short. Then we stash then in some structure, and if a cancel request comes in, we do a much simpler search of the sys.queue() table that's guaranteed to work. JONATHAN ZINGMAN | Manager – TrueCall Data | DIRECT: 510.809.0423 | CELL: 510.823.5638 | www.netscout.com -----Original Message----- From: users-list <users-list-bounces+jonathan.zingman=netscout.com(a)monetdb.org> On Behalf Of Martin Kersten Sent: Tuesday, June 11, 2019 1:56 PM To: users-list(a)monetdb.org Subject: Re: Cancel query This message originated outside of NETSCOUT. Do not click links or open attachments unless you recognize the sender and know the content is safe. On 11/06/2019 22:38, Ying (Jennie) Zhang wrote: > > >> On 11 Jun 2019, at 00:58, Zingman, Jonathan <Jonathan.Zingman(a)netscout.com> wrote: >> >> In the cookbook, you give a good way to find out the status of queries by using the sys.query() call, > > Hai, Jonathan, > > MonetDB doesn’t have a function called sys.query(). Did you mean sys.queue() instead? > >> which returns the list of queries tagged in various ways. Our question is whether there is any way to access, e.g., the tag value. In our application, we need to occasionally cancel queries from our UI. The UI has no direct knowledge of the query. While we might be able to find the query we want to cancel by comparing the query text, our automatically generated queries can be very long, and if there’s some internal reformatting by Monetdb, we might not get an exact match. So is there anything in the handle or elsewhere that we can use to identify the query we want to cancel? > > Correct me if I’m wrong, but I’m not aware of a feature with which you can query the qtag of a query. Some code study will be needed to see if it’s feasible to add such feature, and how. > > You can probably get away with string matching, since MonetDB doesn’t do much/any reformatting next to to-lowercase. Not sure how much text the “query” field of sys.queue() holds. > > In addition, please be aware that > i) sys.queue() is originally develop for debugging purpose… > ii) the minimal cancellation entity is a function, i.e., when MonetDB is busy executing a very long running function, a cancellation will only take effect *after* that function has finished => then MonetDB won’t continue executing the next function. Small correction here, with function we mean a MAL function call. So, if the system is executing a very time consuming join, it will *not* be interrupted, but rather at the end of it the query in which is rans is terminated. > > Regards, > Jennie > >> >> JONATHAN ZINGMAN >> Manager – TrueCall Data >> >> <image001.png> >> 2855 Telegraph Ave, Suite 200 >> Berkeley, CA, 94705-1169 >> >> DIRECT: 510.809.0423 >> CELL: 510.823.5638 >> http://www.netscout.com >> >> >> _______________________________________________ >> users-list mailing list >> users-list(a)monetdb.org >> https://urldefense.proofpoint.com/v2/url?u=https-3A__www.monetdb.org_mailma… > > _______________________________________________ > users-list mailing list > users-list(a)monetdb.org > https://urldefense.proofpoint.com/v2/url?u=https-3A__www.monetdb.org_mailma… > _______________________________________________ users-list mailing list users-list(a)monetdb.org https://urldefense.proofpoint.com/v2/url?u=https-3A__www.monetdb.org_mailma…

2 1

Cancel query
by Zingman, Jonathan 11 Jun '19

11 Jun '19

In the cookbook, you give a good way to find out the status of queries by using the sys.query() call, which returns the list of queries tagged in various ways. Our question is whether there is any way to access, e.g., the tag value. In our application, we need to occasionally cancel queries from our UI. The UI has no direct knowledge of the query. While we might be able to find the query we want to cancel by comparing the query text, our automatically generated queries can be very long, and if there's some internal reformatting by Monetdb, we might not get an exact match. So is there anything in the handle or elsewhere that we can use to identify the query we want to cancel? JONATHAN ZINGMAN Manager - TrueCall Data [logos] 2855 Telegraph Ave, Suite 200 Berkeley, CA, 94705-1169 DIRECT: 510.809.0423 CELL: 510.823.5638 www.netscout.com<http://www.netscout.com/>

3 2

MonetDB - Case Sensitive
by Suresh Devarakonda 11 Jun '19

11 Jun '19

Hi, I am new to the MonetDB and interested to know about its case sensitivity. We noticed that objects are being created in lower case if they are not specified in quotes so it appears to me that default case for objects created in Monet is LOWERCASE. But, we got a requirement that objects should be created in upper case even though they are not specified in quotes in DDL statements. In other words, can we make the default to UPPERCASE for all the objects when they created. Please let us know if that is possible. Your early response is very much appreciated. ~Suresh D

2 1

Installation error on AWS Linux
by Gerald Sangudi 04 Jun '19

04 Jun '19

Hello, I'm trying to install MonetDB on AWS Linux. The first command succeeds: sudo yum install https://dev.monetdb.org/downloads/epel/MonetDB-release-epel.noarch.rpm The second command fails: sudo yum install MonetDB-SQL-server5 MonetDB-client Error: *[ec2-user@ip-172-31-23-149 ~]$ sudo yum install MonetDB-SQL-server5 MonetDB-clientLoaded plugins: extras_suggestions, langpacks, priorities, update-motdhttps://dev.monetdb.org/downloads/ <https://dev.monetdb.org/downloads/>epel/2/x86_64/repodata/repomd.xml: [Errno 14] HTTPS Error 404 - Not FoundTrying other mirror. One of the configured repositories failed (MonetDB 2 - x86_64), and yum doesn't have enough cached data to continue. At this point the only safe thing yum can do is fail. There are a few ways to work "fix" this:* It seems to me there's a reference to epel/2 which is out of date. Should that be epel/7? Is it possible to get an updated rpm? Thanks, Gerald

2 2

REPLICA, CLUSTER
by Ariel Abadi 29 May '19

29 May '19

Hi! Does anybody implemented a REPLICA DB, or a CLUSTER with several nodes for monetdb? I couldnt find too much information in how to implement it. Any information you can provide, will help a lot! Thks in advance Ariel

1 0

MonetDB performance on large table with 20K columns
by Malware Research 16 May '19

16 May '19

Hi, I posted a question on StackOverflow <https://stackoverflow.com/questions/55951563/monetdb-performance-on-a-large…> regarding s specific performance issue. I’m not sure if the right people are monitoring S/O from time to time so I thought I’ll post it here too. Here it is in case you do not want to follow the link: I'm testing MonetDB as a solution for a data-science project. I have a table of 21K columns - all but three are features described as float (32bit) and 6.5M rows (which may or may not become larger, perhaps up to 20M rows). My aim is to use the integrated Python on MonetDB to achieve the ability to train without exporting the data from the DB every time. In addition, queries on specific columns are necessary so the columnar storage can be a significant advantage. I have compiled MonetDB 11.31.13 to gain the embedded Python support. OS is CentOS 7. Storage is not SSD. 48 core server with ~300GB of memory. I created an (unique) index on the table (without analyze). I noticed that when I SELECT * FROM [TABLE_NAME] SAMPLE 50; it takes a long long time to complete. I then tried: SELECT f1, f2, ..., f501 from [TABLE_NAME] SAMPLE 50; SELECT f1, f2, ..., f1001 from [TABLE_NAME] SAMPLE 50; SELECT f1, f2, ..., f2001 from [TABLE_NAME] SAMPLE 50; ... SELECT * from [TABLE_NAME] SAMPLE 50; I ran the queries locally with mclient and used time to measure the amount of time it took and I noticed two things: 1. There is a period where a single core is taking 100% CPU. The more columns the longer it takes to complete. Only when it finishes I can see all cores working, data being consumed, etc... In addition, during that time, the query does not appear in the result of select * from sys.queue(); Eventually, the time needed to get 50 rows from the table was almost 4 hours. 2. The amount of columns is doubled but between each step in the test the amount of time it takes to get a result is tripled. So my questions is: Is this behaviour expected or does it reflect something I did wrong? The data requested from the table should be around 4MB (50 * 21000 * 4Bytes), so this reflects a significant time waiting for such a small amount of data. Help is appreciated! Help is very much appreciated!

2 1

MonetDB mercurial source doesn't compile on macOS Mojave
by Ying Zhang 16 May '19

16 May '19

Dear all, When trying to compile the MonetDB mercurial source on macOS Mojave, it ends with the error: /Users/jennie/monet/Apr2019/common/stream/stream.c:819:10: error: implicit declaration of function 'fdatasync' is invalid in C99 [-Werror,-Wimplicit-function-declaration] && fdatasync(fileno(fp)) < 0 ^ 1 error generated. Does anyone know a fix or workaround for this? Thanks a lot! Jennie

1 1

BPP version problem
by Corbit, Dann 02 May '19

02 May '19

Mdblaunch.bat seems to work: C:\Program Files\MonetDB\MonetDB5>mserver5 --dbpath=dbfarm\voc\ --set mapi_port=50000 # MonetDB 5 server v11.33.3 (Apr2019) # Serving database 'voc', using 12 threads # Compiled for x86_64-pc-winnt/64bit # Found 31.927 GiB available main-memory. # Copyright (c) 1993 - July 2008 CWI. # Copyright (c) August 2008 - 2019 MonetDB B.V., all rights reserved # Visit https://www.monetdb.org/ for further information # Listening for connection requests on mapi:monetdb://127.0.0.1:50000/ # MonetDB/GIS module loaded # MonetDB/SQL module loaded > But neither of these batch files work. C:\Program Files\MonetDB\MonetDB5>MSQLserver.bat MonetDB/Python Disabled: Python 3.7 installation not found. !FATAL: BBPinit: incompatible BBP version: expected 061041, got 061033. This database was probably created by a too old version of MonetDB. Press any key to continue . . . C:\Program Files\MonetDB\MonetDB5>M5server.bat MonetDB/Python Disabled: Python 3.7 installation not found. !FATAL: BBPinit: incompatible BBP version: expected 061041, got 061033. This database was probably created by a too old version of MonetDB. Press any key to continue . . . They complain about the server version. I tried deleting the database folder and restarting. The database was rebuilt. But I still get the errors. How do I fix this error? I do not have real database systems. I only have the voc sample data.

3 3

Trying to link IBM Cognos and MonetDB via ODBC
by SERET Paul 25 Apr '19

25 Apr '19

Dears, I get the following error while trying to link IBM Cognos Analytics software to MonetDB via ODBC (sorry, the error message is in french, I translate below) : [cid:image001.png@01D4FB81.C8C4CA40] QE-DEF-0285 The logon failed QE-DEF-0325 The logon failed for the following reason: RQP-DEF-0068 Unable to connect to at least one database during a multi-database attach to 1 database(s) in: 'testDataSourceConnection' UDA-SQL-0031 Unable to access the 'testDataSourceConnection' database. Check that the connection parameters to the database are configured correctly. For example, ensure that the data source connection contains the signon information, such as a password, to connect to the database. UDA-SQL-0107 A general exception has occurred during the operation 'attach(error code -49)'. The specified driver has not been loaded because of system error 127 : The specified procedure cannot be found. (MonetDB ODBC Driver, C:\Program Files (x86)\MonetDB\MonetDB ODBC Driver\lib\libMonetODBC.dll). The last section of the message sounds strange to me, because the DLL seems to be correctly installed from the most recent ODBC driver installer (2019/01/04) : [cid:image002.png@01D4FB84.B9159040] It looks like there is an issue inside the DLL program. Is there a known issue with this DLL or the driver ? IBM Cognos mentions that the ODBC driver must be compliant with the 3.5 standard. I guess this is the case for MonetDB ODBC driver, isn't it ? Could you please have any clue to get this sorted out ? Thanks in advance, best regards Paul SERET

1 0

How can I get a parameterized bulk version of a scalar C UDF to work in SQL?
by aris 25 Apr '19

25 Apr '19

Hi, Say that I have a family of simple scalar transforming function f(x ,p) => y parameterized by p which are implemented as C UDF's and exposed as SQL UDF's. Is it possible to create a bulk version of this? So the scalar C signature is simplified to something like char* f(int* output, int* input, const char** func_parameter); the mal stuff module m; command f(input:int, func_parameter:str):int address f; and in sql we have CREATE FUNCTION f(input INT, func_param STRING) RETURNS INT EXTERNAL NAME m.f; Now for the bulk version I would like to use in C: char* f(bat* output, bat* input, const char** func_parameter); in MAL: module batm; command f(input:bat[:int], func_parameter:str):bat[:int] address f; in SQL: reuse the scalar SQL signature. But that doesn't seem to work. When I compile this and call the udf in a sql query on an actual column, the plan applies the scalar version on each tuple instead of calling the bulk version once. However if I remove the parametrization, i.e. removing the func_parameter from the signature, the bulk version is correctly called. Aris

1 0

Database crash issue
by Amit Ambekar 12 Apr '19

12 Apr '19

Hi, I am using MonetDB from last 2 years, but never faced such issue till now. I am not able to use database ,it is giving error as below: monetdbd: internal error while starting mserver database has crashed after starting, manual intervention needed, check monetdbd's logfile (merovingian.log) for details' Regards, Amit Ambekar Data Engineer

2 2

Get stuck when trying to copy large amounts of data
by Yinjie Lin 02 Apr '19

02 Apr '19

Hi, I'm trying to copy large amounts of data into MonetDB, but the processes would get stuck in the middle. The following is what I did. I have 6 tables, all merge tables partitioned by range of time. And there are hundreds of data files, each has almost 10 million rows and is of size about 3.5 GB. For each table, I use the following command to copy data: `cat file.txt | xargs -n1 -P1 -I {} sh -c "mclient -p 50000 -s \"copy 10000000 records into tbx from stdin best effort\" - < '{}'" ` where file.txt contains paths of files to be loaded. So 6 mclient connections are created to copy data in parallel. But mclient processes would get stuck after loading more than 200 million rows each table. Then I have to kill the processes and restart the database. The environment is CentOS 7.3, 512 GB RAM. MonetDB is compiled using the latest master branch. So does anyone know why this happened? Does MonetDB have a limit on the size of data one instance can hold? Or if there is something wrong with my practice? Thanks in advance! Yinjie Lin

2 1

Performance issue with count distinct
by AUBERT Simon 29 Mar '19

29 Mar '19

Hello all, Here a simple issue I have with performance : 1/ I create a table A and populate it with a bulk load 2/I create a table B, I populate the table with bulk load 3/I create table C with a CTAS statement. I analyze the statistics and set the primary key 4/I try a "count distinct" type query (on all the table) on the PK Field. It is pretty slow, with a high sollicitation of my hard drive for several dozens of second. the "select count" is however fast. So I have these following questions : 1/why despite statistics the count distinct query is slow? I thought the idea of statistics was to get some metrics about the table? 2/if the field I make a count distinct on is the PK (or the set of field is the PK), it's exactly the same than a count non null. So why the time are not the same? 3/if these issues are design issue or bugs, how can I ask to make it solved? For information : Monetdb 11.31.13 on Windows 10 Best regards, Simon AUBERT [cid:691B9F4089B9F34B8E0C22B39B7117EC@eurprd07.prod.outlook.com] Bourse Maritime - 1, place Lainé - 33 000 Bordeaux Mob : +33(0)6 66 28 52 04

2 1

Program contains errors.:(NONE).multiplex
by Niklas B 29 Mar '19

29 Mar '19

Hi, I'm upgrading from an old MonetDB (11.26.0 source snapshot) to the latest proper release. However, my Python aggregate functions appear to have stopped working (Program contains errors.:(NONE).multiplex): Running mserver5 with: /usr/local/bin/mserver5 --daemon=yes --set embedded_py=true --dbpath=/data/monetdb-farm/ --set gdk_nr_threads=1 --set max_clients=6 Reproduce with: CREATE FUNCTION test_udf("conum" int) RETURNS STRING LANGUAGE PYTHON3 {{ }}; Create table test("id" uuid, "date" DATE, "number" INT); select test_udf("number") from test; Gives: Program contains errors.:(NONE).multiplex I'm running it via Docker (stock Debian) with only MonetDB installed via apt-get. Not sure if this is connected: https://www.monetdb.org/bugzilla/show_bug.cgi?id=6378 <https://www.monetdb.org/bugzilla/show_bug.cgi?id=6378> Anyone having any similar issues? Tried compiling from source (April 2019 bransch) as well, no luck Regards, Niklas

3 2

sys.storage sorted attribute changes for primary key
by Jeremy Norris 29 Mar '19

29 Mar '19

Hello, I have a large table (~700 million rows, ~150 columns). Initially after import, the database is very fast on primary key lookups, however after a few thousand creates/updates/deletes on the database, "sorted" (as reported by sys.storage), flips from true to false. This results in major performance degradation. What are the conditions under which a database can lose it's sorted-ness on the primary key column? eg: sql>select schema, table, column, sorted from sys.storage where schema = 'test1' and table = 'cmdb' and column = 'pk_id'; +---------+-------+--------+--------+ | schema | table | column | sorted | +=========+=======+========+========+ | test1 | cmdb | pk_id | false | +---------+-------+--------+--------+ Thanks, -Jeremy Norris

2 1

Fwd: Question about partition bounds of updatable merge tables
by Yinjie Lin 06 Mar '19

06 Mar '19

Hi, I found that there is a new feature called updatable merge table in the new release (Updatable merge tables with partition predicates <https://www.monetdb.org/blog/updatable-merge-tables>). I think this makes merge table similar to *partitioned table* in other databases. But I found that the upper bound of range partitioning is an inclusive bound, while the upper bound of some databases like MySQL and PostgreSQL is an exclusive bound. So my question is: why is the upper bound declaration designed like this? Will it be possible if MonetDB be consistent with other databases? Besides, currently merge table doesn't support partition pruning, so all child tables will be scanned even if the filter condition contains the partitioning key. Is it possible if this feature will be added one day? Best Regards, Yinjie Lin

1 0

#!ERROR: GDKmove: cannot rename /mnt/g/monet/wal/sql/log to /mnt/g/monet/wal/sql/log.bak
by Michael Jansen 21 Feb '19

21 Feb '19

Hello, I have a crash after about 5 minutes when running in transaction replication mode. After the crash I cannot access the db. I am running: mserver5 --trace --dbpath=/data/monetdb/dbfarm/test_db/ --set gdk_keep_persisted_log_files=1000 --set gdk_logdir=/mnt/g/monet/wal --set sql_debug=4 I am using GlusterFS under /mnt/g to manage the replication. I am running on AWS Linux 2. and I have free space on in both the data and wal directories. (I assume this is enough space): Filesystem Size Used Avail Use% Mounted on 10.0.1.68:/gv0 8.0G 7.4G 669M 92% /mnt/g bash-4.2$ the db only has one test table: sql>create table t (n float, t text) ; operation successful sql>select * from t ; +--------------------------+------+ | n | t | +==========================+======+ | 1.1 | blah | | 1.1 | blah | | 1 | blah | +--------------------------+------+ 3 tuples any help is appreciated! -mj

1 0

Error when installing on Centos 7
by Trevor Andreas 25 Jan '19

25 Jan '19

Hello, I am getting an error trying to install MonetDB following the instructions here: https://dev.monetdb.org/downloads/epel/ The error I am getting is when running: yum install -y MonetDB-SQL-server5 MonetDB-client failure: repodata/repomd.xml from monetdb: [Errno 256] No more mirrors to try. https://dev.monetdb.org/downloads/epel/7/x86_64/repodata/repomd.xml: [Errno 14] HTTPS Error 404 - Not Found Looks like the repodata is missing for Centos 7? It looks to be there for Centos 6 https://dev.monetdb.org/downloads/epel/6/x86_64/repodata/ Thanks! Trevor

2 1

Invalid Credentials error
by AUBERT Simon 17 Jan '19

17 Jan '19

Good morning, I have used Monetdb for a few weeks now and this evening I cannot login after a full system restart (windows+monetdb). Environment : Windows 10 (well, it's my work computer, I don't have a lot of choice) Monetdb 5 server v11.31.11 I start monetdb by using C:\Program Files\MonetDB\MonetDB5\bin\mserver5.exe Then mclient on C:\Program Files\MonetDB\MonetDB5\mclient.bat (and usually it work just fine) I also tried with mclient.exe using the -x command to debug. Error message : C:\Program Files\MonetDB\MonetDB5\bin>mclient -X -d demo user(win32):monetdb password: MAPI = monetdb@localhost:50000 ACTION= mapi_reconnect ERROR = !InvalidCredentialsException:checkCredentials:invalid credentials for user 'monetdb' Of course, I use the default monetdb/monetdb credentials. I have no idea of what happens and moreover I have lost an entire schema (I see that if I connect through Tableau Desktop...). I have done an upgrade on the last version (11.31.13) and same result, impossible to connect. And same result. I spend the entire evening on that a Thank in advance. Simon AUBERT [cid:691B9F4089B9F34B8E0C22B39B7117EC@eurprd07.prod.outlook.com] Bourse Maritime - 1, place Lainé - 33 000 Bordeaux Mob : +33(0)6 66 28 52 04

2 3

table-returning function with subselect in input - no mitosis?
by Roberto Cornacchia 14 Jan '19

14 Jan '19

Hi, * Background: I have a custom string tokenizer that I invoke this way: SELECT * FROM tokenize( (SELECT id, s FROM my_strings) ); The output has of course a 1-to-N rows relation with the input. The function has a bulk implementation, so it takes M rows and it outputs M rows. * The question: No matter how large the input table my_strings is, mitosis does not seem to kick in. Is it not allowed in this case because it cannot assume that my custom function allows to concatenate partial results? Do you see any chance for me to enable mitosis with this scenario? Thanks, Roberto

1 1

Announcement: New Aug2018-SP2 Bugfix release of MonetDB suite
by Sjoerd Mullender 14 Jan '19

14 Jan '19

The MonetDB team at CWI/MonetDB BV is pleased to announce the Aug2018-SP2 bugfix release of the MonetDB suite of programs. More information about MonetDB can be found on our website at <https://www.monetdb.org/>. For details on this release, please see the release notes at <https://www.monetdb.org/Downloads/ReleaseNotes>. As usual, the download location is <https://dev.monetdb.org/downloads/>. Aug 2018-SP2 bugfix release (11.31.13) SQL Frontend * Disabled function sys.getcontent(url). Bug Fixes * 6643: schema name qualifier in create global temporary table json.table6643 ... is not honered * 6645: optimizer treats all the functions with no or constant parameters as constant * 6650: PREPARE SQL statement fails to compile user defined functions with parameter/s * 6651: Multi-column IN clause for subquery produces wrong results * 6653: CREATE TABLE accepts empty table/column name * 6654: Incorrect handling of 'TRUE' in compound select * 6657: Restart sequence with a non atomic sub-query cardinality gives wrong error message * 6660: GRANTing a ROLE is not idempotent * 6662: Concurrency Conflicts Exception string "!transaction is aborted because of concurrency conflicts, will ROLLBACK instead" not shown on prompt. * 6664: mserver5 crash: infinite recursive happens at rel_bin.c:489 * 6665: Creation of serial types does not accept negative numbers * 6666: COPY INTO from .. LOCKED doubles input data * 6668: The SAMPLE key word doesn't work in a subquery. * 6669: COPY [xxx RECORDS] INTO foo FROM STDIN ... doesn't work without specifying nr of to be copied records * 6672: SQLGetData with SQL_C_WCHAR string truncation and invalid StrLen_or_Ind value

1 0

Re: Performance degradation
by Jeremy Norris 12 Jan '19

12 Jan '19

Here's some additional observations on this puzzling issue: The slow table is "cmdb", the normal table is "cmdb2". Both have the same number of rows according to "SELECT count(*)" (73553310). However, sys.statistics shows something a bit different: sql>SELECT x.name, count(*) AS columns_with, z.count FROM sys.tables AS x, sys.columns AS y, sys.statistics AS z WHERE x.id = y.table_id AND y.id = z.column_id GROUP BY x.name, z.count; +-------+--------------+----------+ | name | columns_with | count | +=======+==============+==========+ | cmdb | 8 | 73553310 | | cmdb | 145 | 73553311 | | cmdb2 | 153 | 73553310 | +-------+--------------+----------+ 3 tuples sql> For the slow table, "cmdb", there are 145/153 columns that have an extra "row". This row count mismatch causes it to skip parallelization optimizations in "monetdb5/optimizer/opt_mitosis.c", line 212. A few related questions: 1) In the columnar model, is is valid for conceptual tuples to exist which don't have values for each column like this? (This seems strange as now there's a difference between an explicit NULL and "not existing"). How might this situation have happened? (Only transactional full-row inserts were applied to this DB). Can this be repaired? 2) Is the conditional in "monetdb5/optimizer/opt_mitosis.c", line 212 sufficiently robust? I may be misunderstanding the full context of this line, but it seems like we could still benefit from splitting/parallelization when r < rowcnt, no? Thanks, -Jeremy Norris

1 0

Performance degradation
by Herman Knief 10 Jan '19

10 Jan '19

Greetings, Running MonetDB v11.31.11 (Aug2018-SP1), we ran into an issue where queries that have been running quite well, suddenly slowed by about 5x. Looking at the explain, it seems we hit a condition that is preventing mitosis? Copying the tables related to the query seems to restore performance to the previous level, but would be very curious to understand what happened here. What metrics or stats should we be looking at? Changes we see in the plan are as follows: Bad example: sql.tid(X_4:int, "waltest":str, "cmdb":str); Good example: sql.tid(X_4:int, "waltest":str, "cmdb2":str, 0:int, 32:int); Looking at the storage info, the sizes (heap and column), type width and counts look the same between the two tables. Any thoughts or suggestions would be appreciated. - Herman -- Please excuse typos... I can't type.

1 0

Embedded Python module is not available in recent builds of Windows version
by Kevin Chen 21 Dec '18

21 Dec '18

Hi, I upgraded the MonetDB windows version to 2018Aug and found that the embedded Python module is no longer working. Even the startup script with “embedded_py=true”, the module was not loaded. There was no message showing “MonetDB/Python2 module loaded”. All my Python functions are failed. Please help. Thanks. Kevin

2 1

!FATAL: SQL type wrd missing
by Ariel Abadi 18 Dec '18

18 Dec '18

Hi! Please I need an urgent solution (if there is one) I have upgraded a database to the latest version, and when Im trying to start the DB, there is an error: 2018-12-18 10:58:06 ERR DB[29818]: !FATAL: SQL type wrd missing Can someone help me with any clue!???? Thks in advance Ariel

2 3