I have an application using HSQL 1.7.2 that contains table definitions with default values of NaN for some of the DOUBLE columns. After issuing, with 1.7.2,
SET SCRIPTFORMAT TEXT
SHUTDOWN SCRIPT
the following appears in the database.script file:
CREATE TABLE ... DOUBLE DEFAULT 'NaN' NOT NULL ...
I'm attempting to upgrade to HSQL 2.3.2, but I first have to upgrade to 1.8. I'm finding that after the upgrade to 1.8, the database.script file has:
CREATE TABLE ... DOUBLE DEFAULT 0E0/0E0 NOT NULL ...
When I open this database with HSQL 2.3.2, I get "SQLException: error in script file line ..." on the line at which "/" first appears. The traceback includes "Caused by: org.hsqldb.HsqlException: unexpected token: /".
I've messed around a lot with double_nan=false, without success.
Does anybody have any suggestions for me?
Support for NaN as default value was dropped in version 2.x. This will be restored in the next update.
For the time being, change the default to NULL and add a TRIGGER on the table to convert NULL entries to NaN.
The property hsqldb.double_nan=false must be specified or the statement SET DATABASE SQL DOUBLE NAN FALSE executed for the database to accept the NaN data values .
Related
After launch the application, I get an error:
Caused by: liquibase.exception.ValidationFailedException: Validation Failed: 1 change sets checksum was: 8:2b2936713e8d9aea052c3122fd81faec but now it is:
8:ed8f7550fdd9809f4f6bf0f2d83dbbd8
The error points to such a table:
create table car (
id bigint not null auto_increment PRIMARY KEY,
name varchar(255) not null,
category varchar(255) not null
);
I read about this error and it was pointed out to use the mvn liquibase:clearCheckSums command but in the terminal I get the error: Error: -classpath requires class path specification (I run the command in the project folder)
The error indicates that a changelog that already run now is running with some changes. If you change something on an already run changelog then the checksum changes and this error is completely normal.
Check your databasechangelog table for the mentioned checksum to verify the changelog that crashes. The solutions are usually not to change the checksum of an already run changelog. If you can drop this changelog from the database and run it again will work fine. Sometimes the same changelog crashes without changing anything. This happening to me all the time because of the line seperator. For example, on IntleliJ you can change it from here:
Check what line separator do you need by testing :)
during the upgrade process from 7.6 to 8.7 the database Analyzer fails in the following statement:
Error:
Database update failed
Error: Specified key was too long; max key length is 1000 bytes
can you help?
i found the solution.
drop index lookup_string on sys_refindex;
then you can run database analyzer in to steps:
ALTER TABLE ....
CREATE INDEX ....
but this is done by database analyzer in upgrade wizard
To transfer data from Ms sql server 2008 to Snowflake I used talend , but every time I get error as
java.io.IOException: net.snowflake.client.loader.Loader$ConnectionError: State: CREATE_TEMP_TABLE, SQL compilation error: error line 1 at position 68
invalid identifier '"columnname"'
at org.talend.components.snowflake.runtime.SnowflakeWriter.close(SnowflakeWriter.java:397)
at org.talend.components.snowflake.runtime.SnowflakeWriter.close(SnowflakeWriter.java:52)
at local_project.load_jobnotes_0_1.Load_Jobnotes.tMSSqlInput_1Process(Load_Jobnotes.java:2684)
at local_project.load_jobnotes_0_1.Load_Jobnotes.runJobInTOS(Load_Jobnotes.java:3435)
at local_project.load_jobnotes_0_1.Load_Jobnotes.main(Load_Jobnotes.java:2978)
Caused by: net.snowflake.client.loader.Loader$ConnectionError: State: CREATE_TEMP_TABLE, SQL compilation error: error line 1 at position 68
invalid identifier '"ID"'
at net.snowflake.client.loader.ProcessQueue.run(ProcessQueue.java:349)
at java.lang.Thread.run(Thread.java:748)
Caused by: net.snowflake.client.jdbc.SnowflakeSQLException: SQL compilation error: error line 1 at position 68
The column does exist in my Snowflake DB still I get error as column does not exist
On analysing what query Talend executing in snowflake I found that It tries to create a temporary table to store data but in doing so it selects all column from table between “ ” double quotes and hence error comes as invalid identifier '"columnname"'
If I execute the same query manually without double quotes its works fine , can you please let us know what is workaround of this issue
Query executed by talend in snowflake for your reference
CREATE TEMPORARY TABLE "Tablename_20171024_115736_814_1"
AS SELECT "column1","column2","column3"
FROM "database"."schema"."table" WHERE FALSE
The issue is most likely due to a case mismatch between the object names in Snowflake and what is being sent through the connector. On the Snowflake side, all object names are stored as UPPER CASE. Suggest you try passing COLUMN1, COLUMN2, etc and see if that works.
You can also try setting the QUOTED_IDENTIFIERS_IGNORE_CASE to true, it might help.
I found that this issue is due to mixed case database or schema names not properly being applied by Talend. I discover a hack by updating the Snowflake connector role parameter and added something such as this screenshot:
I'm working with: seam 2.2.2 + hibernate + richfaces + jboss 5.1 + postgreSQL
I have an module which needs to load some data from the database. Easy. The problem is, on development it works fine, 100%, but when I deploy on my production server and try to get the data, an error rise:
could not read column value from result set: fechahor9_504_; Bad value for type timestamp : [C#122e5cf
SQL Error: 0, SQLState: 22007
Bad value for type timestamp : [C#122e5cf
javax.persistence.PersistenceException: org.hibernate.exception.DataException: could not execute query
[more errors]
Caused by: org.postgresql.util.PSQLException: Bad value for type timestamp : [C#122e5cf
at org.postgresql.jdbc2.TimestampUtils.loadCalendar(TimestampUtils.java:232)
[more errors]
Caused by: java.lang.NumberFormatException: Trailing junk on timestamp: ''
at org.postgresql.jdbc2.TimestampUtils.loadCalendar(TimestampUtils.java:226)
I can't understand why it works on my machine (development) and why not on production. Any clues? Anyone gone through the same problem? Is exactly the same compilation
Stefano Travelli was right. I was checking the jBoss on production and there was an old jdbc driver on [jboss_dir]/common/lib from an old jwebstart application (not developed by me). Deleted that jdbc and it works fine. I should check if the old application is still needed and if so, check if it still works without the jdbc being there or with an upgraded version.
Not sure what the driver story is ..
but the problem for me show up when JDBC tries to parse bigint from the DB to
myOjbect.setDate(Date date){...}
the other "JDBC friendly" is ignorred for some reason .
myOjbect.setDate(long date){...}
So .. removing the Date setter and leaving a long one resolves the problem.
This is a Big workaround .. but may help someone out there :)
Whenever I connect to HSQLDB from my application deployed on JBoss 5.1, it throws exception as : Caused by: org.jboss.resource.JBossResourceException: Could not create connection; - nested throwable: (java.sql.SQLException: error in script file line: 1 Unexpected token UNIQUE, requires COLLATION in statement [SET DATABASE UNIQUE]) .
My HSQLDB script file reads something as below :
`SET DATABASE UNIQUE NAME HSQLDB2E0BAD63B3
SET DATABASE GC 0
SET DATABASE DEFAULT RESULT MEMORY ROWS 0
SET DATABASE EVENT LOG LEVEL 0
.....`
Does anyone have idea what thos exception means or should I change something in my HSQLDB configuration?
Regards,
Satya
Your database files are created with version 2.x, but the version of database engine running on JBoss 5.1 is 1.8.x. You should be able to replace the hsqldb.jar in the JBoss configuration with a new version.