Rocket® Vertica

test question

Hi,Is there a way to explicitly lock a table in Vertica before writing to it? The intent is to limit the write to one session at a time. I know about select for update, but what if the record does not even exist.. Also if you can explain how Serializable isloation level affects other sessions if at all it does (which I believe it does not) and what is the real use of it.Thanks

I'm trying to use vertica with Golang over ODBC. I can connect OK and am getting an error of "Syntax error at or near "\" at character 31"  for the query "select * from public.value_ref" - character 31 would be after the last letter of the table. No variations to the query have resolved error. This does work through JDBC and isql over ODBC with this same DSN.

Hello,I continuously load data and execute queries on them in Vertica and I rely on sysquery tables to perform regular house-keeping. So I launch a lot of requests towards the v_monitor tables, especially the v_monitor schema.partition_status.These requests looks quite slow to me.I checked they have enough memory to perform ( I tried to increase sysquery memorysize or I also checked that when default sysquery memorysize is used, the general pool is free enough to accept that sysquery borrows from there).As the requests below shows it, these sysquery requests are queued and a lot of them waits for more than 5 sec in the queue.=> select * from(SELECT pool_name, queue_entry_timestamp, acquisition_timestamp, (acquisition_timestamp-queue_entry_timestamp) AS queue_wait, node_name FROM V_MONITOR.RESOURCE_ACQUISITIONS WHERE node_name ILIKE '%node000%' and pool_name LIKE '%query%') foo where queue_wait >= '00:00:05'::INTERVAL; pool_name |     queue_entry_timestamp     |     acquisition_ti

The Vertica version is v7.0.1 runs in 3 nodes.I started uploading large files of data to the vertica (files of 100,000 records each, about 2-3 per node per minute).I see that after a little time, all my available threads get stuck on actions with vertica (copy or select  ,vsql selects not responding either). Vertica does not return a timeout so my whole server is stuck.I do not understand why vertica gets stuck. I can see in its logs no errors. This happens gradually on all nodes until nothing is working anymore (besides internal Vertica processes such as tupple mover etc.)

Can I use multiple different R scripts to populate a single Vertica UDF Library? Or does all my code has to be in a single .R file?

I have a SQL statement that is dynamically created; that is, it is comprised of a static SQL string with some dynamic value appended to it. How do I go about executing this statement? Is there a command for this? This would be equivalent to "EXEC sp_executesql SqlString" in SQLServer.Also, what is the syntax for looping through a result set? Is there such a command? If so, can you please share an example of this?What i'm trying to do is: Run a SELECT statementLoop through the result set to create the SQL stringRun the SQL string

i just define "outtype=c("any")" , and when i create library, it prompt me to define a outtype call back function.

Hi,I have a three nodes VERTICAcluster with Ksafe=1.  When I create a table with “create table test_proj as  select * from all_tables” I got two projections with name “test_proj_b0” and “test_proj_b1”.  I think one is primary projection another is buddy projection, right? I want to know which one is primary projection which one is buddy projection?if I run a query against  this table, which one will serve the query?

I would like to access tables in Vertica such as the disk_storage table using the ODBC interface.  The user level DB's are detailed in the v_catalog.columns however I cannot seem to see any detailed about the v_catalog tables, possibly due to their being virtual.So the questions is if I want to use ODBC and do a bind of C data types to the statement results how do I determine what the column data types will be in the disk_storage table for example ?             std::string statement("SELECT * FROM disk_storage WHERE storage_usage ILIKE '%data%' ORDER BY rank;"); ...             ret = SQLExecDirect( hdlStmt, (SQLCHAR*)statement.c_str(), SQL_NTS );             ret = SQLBindCol(hdlStmt, 1, SQL_C_???????, ....;What is the secret way to determine what  SQL_C_???????,would be for the system tables?Thanks,Karl

Hi. Does the distributed R support the igraph package? Regards Sanjay

Hi Folks. I've run you Database Designer, but my queries are now about 25% slower. I posted this at why does running the database designer result in poorer performance?, but I've decided to revert to the default superprojections, which worked better for me. How do I do that? Thanks in advance.

Hello,One of the nodes on our 3 node AWS cluster started to shut down with the following message: 2014-08-03 22:26:54.328 EEcmdq:0x7f32f0cec950 [Main] <PANIC> Received fatal signal SIGSEGV. 2014-08-03 22:26:54.328 EEcmdq:0x7f32f0cec950 [Main] <PANIC> Info: si_code: 2, si_pid: -2087288392, si_uid: 32564, si_addr: 0x7f34839681b8This has been happening the last 2 days out of the sudden, happened multiple times and only to the same node. Can someone help me to troubleshoot this?Thank you,Tibor

Hi all, I am getting prepared to get certified in Vertica and I am having doubts regarding these 4 questions particularly....the answers that I think are correct are marked, but I am not entirely sure, could someone please advise... it would be greatly appreciated. thanks in advance.  1.    Which statements are true about a prejoin projection? (Select three.) A) A prejoin projection enforces primary and foreign key constraints.    -correct A prejoin projection does not increase the storage footprint. C) A prejoin projection slows down data loading. D) A prejoin projection can process any type of join. E) A prejoin projection adds no limitations to the tables in the join.    -correct F) A prejoin projection improves query run time performance.    -correct 2.    What is a benefit of having identically-sorted buddy projections? A) improved storage usage, since Vertica shares data files more efficient use of disk since the data is only on one node C) b

After installing Community Edition of Vertica (see version in subject) on my CentOS installation, I can't set the Restart Policy for the db I created as it returns the error:Unable to restart policy until cluster is fully reachable.The following communication failed:127.0.0.1 from 127.0.0.1It is a single-node cluster with disabled iptables .SSH passwordless has been already set for dbadmin user.How can I proceed ?Thank youIgor

Hi,Can somebody explain the exact meaning of the parameter Moveout Size of the Tuple Mover config ?And also give me concrete cases when we can configure it differently than the default value which is zero.Thanks in advance,Br

My current data usage looks like: dev=> select * from disk_storage;      node_name      |                    storage_path                    | storage_usage | rank | throughput | latency | storage_status | disk_block_size_bytes | disk_space_used_blocks | disk_space_used_mb | disk_space_free_blocks | disk_space_free_mb | disk_space_free_percent---------------------+----------------------------------------------------+---------------+------+------------+---------+----------------+-----------------------+------------------------+--------------------+------------------------+--------------------+------------------------- v_armdevdb_node0001 | /data/ARMDEVDB/v_armdevdb_node0001_catalog/Catalog | CATALOG       |    0 |          0 |       0 | Active         |                  4096 |               29380292 |             114766 |                1104914 |               4316 | 3% v_armdevdb_node0001 | /data/ARMDEVDB/v_armdevdb_node0001_data            | DATA,TEMP     |    0 |          0 |     

I am trying to convert our oracle database to vertica. I found solution from community and documentation regarding loading data and support most of RDBMS. Reading on ROS functionality want to know is Dimension tables specially Type 1 & Type 2 where it logical delete and update the same record will be possible in Vertica ?Also Vertica documentation states support partition but for Fact tables, I have very large dimension table which is partition is oracle need to partition   will be ok to do in Vertica or will be an issue ?can you please share your thoughts.

The 7.1.x documentation online states that we cannot use the 6.1.x or 7.0.x drivers with Vertica 7.1.x. This would be a significant barrier to allowing us to upgrade as we have applications that access Vertica periodically and are using both the 6.1.x drivers as well as the 7.0.x drivers.  Knowing that a strict cutover requirement would be a very resource (person / Project management / upper mgmt) heavy task for any organizaation I am guessing that the documentation is just not correct.  Would someone from Vertica please verify this documentation?  Can we use the older drivers while we upgrade Vertica to 7.1.x and then roll out upgrades to the client drivers in our applications along with their next natural revision cycle?http://my.vertica.com/docs/7.1.x/HTML/index.htm#Authoring/SupportedPlatforms/ClientDrivers.htm%3FToc...Thanks.

Just doing some initial tests setting up a LAP, it doesn't seem as though the actual LAP appears in the query plan. Is this expected behavior as the anchor table is used, or is the setup below incorrect?CREATE PROJECTION online_sales.online_sales_fact_lap_anchor ( transaction_type, warehouse_key, gross_profit_dollar_amount ) AS SELECT transaction_type, warehouse_key, gross_profit_dollar_amount FROM online_sales.online_sales_fact SEGMENTED BY HASH(online_sales_fact.transaction_type, online_sales_fact.warehouse_key) ALL NODES KSAFE 1; SELECT MARK_DESIGN_KSAFE(1); SELECT REFRESH('online_sales.online_sales_fact'); CREATE PROJECTION online_sales.online_sales_fact_lap AS SELECT transaction_type, warehouse_key, SUM(gross_profit_dollar_amount) AS sum_gross_profit_dollar_amount FROM online_sales.online_sales_fact GROUP BY transaction_type, warehouse_key; SELECT REFRESH('online_

Hi we are looking for a vertica expert here at our companyWe need a professional who has maybe done one or two implementations that would assess our integration plans, environment setups and perform optimization. This would be in the West Coast office

I know thatvertica has a big powerfull in analytics, with his own functions and givingtools to programme your own code in R and put inside vertica for example.   But could you access this functions since BI tools as QlickView or Pentaho, I mean, I know that you can connect Vertica with this tools, but I’m not sure if you can execute selects with analytics parameters directly.   Example : SELECT slice_time, symbol, TS_FIRST_VALUE (bid,'LINEAR') AS bid FROM tick_store4 TIMESERIES slice_time AS '2 seconds' OVER (PARTITION BY symbol ORDER BY ts); And your "R" Funcions? Regards

Hello, HP Vertica Place User, We are pleased to let you know that HP Vertica Place 0.6.1 is now available for download on the HP Vertica Marketplace! To access, please visit the Marketplace, and click on the Innovations tab to see the HP Vertica Place download. This iteration of HP Vertica Place will work on all versions of Vertica 7.0.x. Updates include: Significantly improved performance for creating indexes Intersections of points and polygons are now 100% accurate Significantly improved performance in finding intersections of points with polygons. Please see the Release Notes included in the download for changes to the affected functions. HP Vertica Place is a part of our Innovations program, and should be used in test environments only. You can find more information about HP Vertica Place by visiting our post in the HP Vertica Community and by watching a webinar that covers both HP Vertica Place and HP Vertica Pulse. We look forward to your feedback! Please post q

I noticed a new column (ORDER_BY_TYPE) in the 7.1 documentation. Is a descending sort order coming to projections in a future release or is there an alternative meaning to the column (related to Top-K LAP)?

is the Vertica Big Data Conference free for the public?