phoenix-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Amit Shah <amits...@gmail.com>
Subject Re: Missing Rows In Table After Bulk Load
Date Mon, 11 Apr 2016 09:59:12 GMT
Yes, It looks like they were not unique and hence the reduction in count.

Thanks!

On Fri, Apr 8, 2016 at 9:46 PM, Steve Terrell <sterrell@oculus360.us> wrote:

> Are the primary keys in the .csv file are all unique? (no rows overwriting
> other rows)
>
> On Fri, Apr 8, 2016 at 10:21 AM, Amit Shah <amits.84@gmail.com> wrote:
>
>> Hi,
>>
>> I am using phoenix 4.6 and hbase 1.0. After bulk loading 10 mil records
>> into a table using the psql.py utility, I tried querying the table using
>> the sqlline.py utility through a select count(*) query. I see only 0.1
>> million records.
>>
>> What could be missing?
>>
>> The psql.py logs are
>>
>> python psql.py localhost -t TRANSACTIONS_TEST
>> ../examples/Transactions_big.csv
>> csv columns from database.
>> Table row timestamp column position: -1
>> Table name:  SYSTEM.CATALOG
>> CSV Upsert complete. 10000000 rows upserted
>> Time: 4679.317 sec(s)
>>
>>
>> 0: jdbc:phoenix:localhost> select count(*) from TRANSACTIONS_TEST;
>> Table row timestamp column position: -1
>> Table name:  TRANSACTIONS_TEST
>> Table row timestamp column position: -1
>> Table name:  SYSTEM.CATALOG
>> +------------------------------------------+
>> |                 COUNT(1)                 |
>> +------------------------------------------+
>> | 184402                                   |
>> +------------------------------------------+
>> 1 row selected (2.173 seconds)
>>
>> Thanks,
>> Amit
>>
>
>

Mime
View raw message