phoenix-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Sudhir Babu Pothineni <>
Subject Re: Scan phoenix created columns, hbase
Date Mon, 05 Jun 2017 21:41:58 GMT
Cheyenne, We need to read the SYSTEM:CATALOG to get the column mapping.

James/Samarth, I am in the same boat, more pointers into the code will be helpful, after upgrading
to 4.10, our local API code to Phoenix broke because of column mapping.

For now I am parsing the SYSTEM:CATALOG, is there any easy way of doing it?


Sent from my iPhone

> On Jun 5, 2017, at 3:21 PM, Cheyenne Forbes <> wrote:
> So there is no Java function to turn "\x80\x0F" back into "fname" if I'm using CellUtil.cloneQualifier(cell)?
> Regards,
> Cheyenne O. Forbes
>> On Mon, Jun 5, 2017 at 3:13 PM, Samarth Jain <> wrote:
>> Cheyene, with Phoenix 4.10, column mapping feature is enabled by default which means
the column names declared in the Phoenix schema are going to be different from the column
qualifiers in hbase. If you would like to disabled column mapping, set COLUMN_ENCODED_BYTES=NONE
property in your ddl.
>>> On Mon, Jun 5, 2017 at 1:09 PM, Cheyenne Forbes <>
>>> Can anyone please help?
>>>> On Sat, Jun 3, 2017 at 8:51 PM, Cheyenne Forbes <>
>>>> I am doing some analytics which require me to scan through a phoenix created
table with hbase instead of a phoenix select query.
>>>> I created a column with the name 'fname' but a scan with hbase shell shows
>>>> \x00\x00Z\xF3\x10z@\x14    column=PERSONAL:\x80\x0F, timestamp=1496360923816,
>>>> How can I scan using column names if for example the name i gave is "fname"
but in hbase I it is "\x80\x0F"?
>>>> Regards,
>>>> Cheyenne O. Forbes

View raw message