I am a newbie in HIVE. Now I have a .csv file containing the data that I want to append to a table in the HIVE database. But when I run the command: load data inpath /path/to/file into table T1 and wait for a long time, an exception is thrown: "Error in acquiring locks: Locks on the underlying objects cannot be acquired. retry after some time". Those SQL queries all need a very long time (maybe 1~2 hours) to return the results.
I check the locks in this table T1 via show locks T1 extended, and I find some queries are running on the same table:
T1: SHARED
LOCK_QUERYID:hive_20211203..... (omitted)
LOCK_TIME:1638512103766
LOCK_MODE:IMPLICIT
LOCK_QUERYSTRING: ... (a SQL query involved in table T1)
... (other queries like above)
I look up the configurations of HIVE:
SELECT VERSION() -> 2.1.1-cdh6.2.1 r0a45... (omitted)
SET hive.lock.numretries -> 100
SET hive.unlock.numretries -> 10
SET hive.lock.sleep.between.retries -> 60s
SET hive.txn.manager -> org.apache.hadoop.hive.ql.lockmgr.DummyTxnManager
SET hive.support.concurrency -> true
The table is an internal table (MANAGED_TABLE), and not a partition table.
My question is how can I successfully load data into the table while the queries are running on the same table.
If the table is not in transactional mode, you can put files into table location using hdfs dfs -put command:
hdfs dfs -put localfilepath/* /your/table/location/
It can affect running queries, so, do it on your own risk.
Related
I have a table in BigQuery with 100 columns. Now I want to append more rows to it via Transfer but the new CSV has only 99 columns. How should I proceed with this?
I tried creating a schema and adding that column as NULLABLE but it didn't work
I am presuming your CSV file is stored in GCS Bucket and trying to use BQ Data Transfer service to load data periodically by scheduling it.
You can not directly Load/Append data into BQ Table due to schema mismatch.
But as an alternative, create a Staging table named staging_table_csv with 99 columns and Schedule a Data transfer service to load CSV to this table on Overwrite mode.
Now write a query to Append the contents of this staging table staging_table_csv to the target BQ table.
Query might look like this:
#standardSQL
INSERT INTO `project.dataset.target_table`
SELECT
*,
<DEFAULT_VALUE> AS COL100
FROM
`project.dataset.staging_table_csv`
Now schedule this query to run after the staging table is loaded
Make sure to keep a buffer between the Staging table load and the Target Table load. You can perform trials to find a suitable buffer.
For eg: If Transfer is scheduled at 12:00, Schedule Target Table load Query t 12:05 or 12:10
Note: Creating an extra Staging table would incur storage costs but
since it is overwritten for each load, historical data cost is not
incurred
I am currently setting up a simple NiFi flow that reads from a RDBMS source and writes to a Hive sink. The flow works as expected until the PuHiveSql processor, which is running extremely slow. It inserts one record every minute approximately.
Currently is setup as a standalone instance running on one node.
The logs showing the insert every 1 minute approx:
(INSERT INTO customer (id, name, address) VALUES (x, x, x))
Any ideas about why this may be? Improvements to try?
Thanks in advance
Inserting one record at a time into Hive will result extreme slowness.
As your doing regular insert into hive table:
Change your flow:
QueryDatabaseTable
PutHDFS
Then create Hive avro table on top of HDFS directory where you have stored the data.
(or)
QueryDatabaseTable
ConvertAvroToORC //incase if you need to store data in orc format
PutHDFS
Then create Hive orc table on top of HDFS directory where you have stored the data.
Are you poshing one record at time? if so may use the merge record process to create batches before pushing into HiveQL,
It is recommended to batch into 100 records :
See here: https://nifi.apache.org/docs/nifi-docs/components/org.apache.nifi/nifi-hive-nar/1.5.0/org.apache.nifi.processors.hive.PutHiveQL/
Batch Size | 100 | The preferred number of FlowFiles to put to the database in a single transaction
Use the MergeRecord process and set the number of records or/and timeout, it should speed-up considerably
How can I programatically find all Impala tables that need INVALIDATE METADATA statement (because they were created in Hive, but not yet known to Impala) or REFRESH (because column added, datafile added, etc.)?
Invalidate Metadata:
As a workaround, create a shell script to do the below steps.
Using beeline, connect to a particular database and run show tables statement and save output data to a file.
Using impala-shell, connect to the same particular database and run show tables statement and save output data to another file.
Now compare both the file to remove the duplicates and get the unique tables list from the first file which is a list of tables which are only in hive but not in impala.
Note:
a. Instead of a particular database each at a time in 1 and 2 steps, you can loop over all databases and save the output to a file. Inside the loop itself, you can redirect and append the output files to another final output file with data in some format like database.table or database_table to get all tables from all databases into a single file. Finally, follow step 3.
b. The unique tables from the second output file after removing duplicates will be tables that are deleted in hive and invalidate metadata needs to be run in impala to remove them from the impala list.
c. Rename of a table in impala can be recognized by hive but vice-versa is not possible and invalidate metadata should be run for both old and new table names to remove and add respectively in impala. This applies to most operations not just rename of table.
Refresh:
Consider a text format table with 2 columns and 1 row data.
Now suppose, a third column is added to that table in the beeline.
select * from table; ---gives 3 columns in beeline and 2 columns in impala since refresh is not run on impala for this table.
If we run compute stats in impala before running refresh in this case, then that newly added column from the beeline will be removed from the table schema in hive as well.
select * from table; ---gives 2 columns in beeline and 2 columns in impala since compute stats from impala deleted the extra column metadata of table although data resides in hdfs for that column. This might cause parsing issues in impala if the column is added somewhere in the middle or front instead of ending.
So it is advised to run REFRESH table name in impala right after adding a new column or doing any modifications in beeline for an existing table to not lose table schema as explained in the above scenario.
refresh table; ---Right after modification in hive run refresh in impala.
select * from table; ---gives 3 columns in beeline and 3 columns in impala since refresh is run before compute stats in impala.
When a session is loading the table on a different partition using the SQLLDR to spool one flat file , we see other sessions on the same table starts long running and are failing after a long time with the error
"ORA-08103: object no longer exists"
We did some initial check for this.
Our table is partitioned and we are using partition key to load from the flat file.
Table's Object id is not changing during that period. Only the partition's Id is changing because we are doing a truncate partition before loading.
We have no global index on it,all indexes are local.
Some flat files are too huge and sometime the AWR report shows complete I/O system consumed.
I'm learning Hadoop/Big data technologies. I would like to ingest data in bulk into hive. I started working with a simple CSV file and when I tried to use INSERT command to load each record by record, one record insertion itself took around 1 minute. When I put the file into HDFS and then used the LOAD command, it was instantaneous since it just copies the file into hive's warehouse. I just want to know what are the trade offs that one have to face when they opt in towards LOAD instead of INSERT.
Load- Hive does not do any transformation while loading data into tables. Load operations are currently pure copy/move operations that move datafiles into locations corresponding to Hive tables.
Insert-Query Results can be inserted into tables by using the insert clause and which in turn runs the map reduce jobs.So it takes some time to execute.
In case if you want to optimize/tune the insert statements.Below are some techniques:
1.Setting the execution Engine in hive-site.xml to Tez(if its already installed)
set hive.execution.engine=tez;
2.USE ORCFILE
CREATE TABLE A_ORC (
customerID int, name string, age int, address string
) STORED AS ORC tblproperties (“orc.compress" = “SNAPPY”);
INSERT INTO TABLE A_ORC SELECT * FROM A;
3. Concurrent job runs in hive can save the overall job running time .To achieve that hive-default.xml,below config needs to be changed:
set hive.exec.parallel=true;
set hive.exec.parallel.thread.number=<your value>;
For more info,you can visit http://hortonworks.com/blog/5-ways-make-hive-queries-run-faster/
Hope this helps.