Re: [GENERAL] using hstore to store documents

Поиск
Список
Период
Сортировка
От Rita
Тема Re: [GENERAL] using hstore to store documents
Дата
Msg-id CAOF-KfgO1gQF=ZSWC+X84iXyQhe99NAo+=ztaSUoCfg9eUtbvQ@mail.gmail.com
обсуждение исходный текст
Ответ на Re: [GENERAL] using hstore to store documents  (Adrian Klaver <adrian.klaver@aklaver.com>)
Список pgsql-general
I want to over write the data because I don't need to keep its history. 
Yes, the 5-6 table will be the subset of the data.  The subset of the data will all come from xpath. I don't want to keep doing an xpath query when a client requests the data.



On Sun, Jan 29, 2017 at 5:42 PM, Adrian Klaver <adrian.klaver@aklaver.com> wrote:
On 01/28/2017 05:57 PM, Rita wrote:
sorry for the late reply.

My table schema is very simple

DROP TABLE xmltest;





create table xmltest(


  id serial,        -- dont really need the serial

Maybe not a serial id, but a Primary Key of some sort would help with what you say you want to do below, I think.



data xml NOT null


);











INSERT INTO xmltest (data, id) VALUES ('


<attendee>


 <bio>


 <name>John Doe</name>


 <birthYear>1986</birthYear>


 </bio>


 <languages>


 <lang level="5">php</lang>


 <lang level="4">python</lang>


 <lang level="2">java</lang>


 </languages>


</attendee>', 1);

I really don't need the serial but every 30 seconds or so I plan to
overwrite the data portion. so, I suppose I wanted a simple key/value
thats where my orignal question stemmed from.

Why do you want to overwrite the data if you plan to refer to it below?


After xmltest has been populated, I can run xpath and unest to get my
data into a row but I would like to store that result in another table,
I am guessing I should look into triggers for something like that?
Eventually, I plan to have 5-6 downstream tables which will have xmltest

Why 5-6 tables?

Are they each holding some subset of data?

A schematic representation of what you are thinking of doing would help with developing an answer to your question.

as my head. The application will be accessing the downstream tables and
rarely be touching xmltest (head table).




On Thu, Jan 26, 2017 at 10:38 AM, David G. Johnston
<david.g.johnston@gmail.com <mailto:david.g.johnston@gmail.com>> wrote:

    On Thu, Jan 26, 2017 at 5:37 AM, Rita <rmorgan466@gmail.com
    <mailto:rmorgan466@gmail.com>>wrote:

        of course, sorry for being vague.

        I have an external process generating a XML file (every 30 secs)
        which is about 10MB. I would like to store the file as XML type
        for me to query using xpath. I plan to query it every few
        seconds by few hundred clients. so, it maybe easier for me
        create a separate table of my xpath results and have clients
        query that table (xpath can be expensive).


    ​If the XML being generated has a fixed structure/schema I
    personally would treat the XML as a serialization format and
    de-serialize and store it in a database as one or more relationally
    linked tables.​  If you have to deal with the possibility of dynamic
    structure I would still try to put the fixed items into individual
    columns and then and then any dynamic items could be stuffed into an
    hstore typed table.

    My answer to your stated question is: what happened when you tried
    doing that?  Documentation and a bit of experimentation goes a long
    ways in learning.

    David J.




--
--- Get your facts first, then you can distort them as you please.--


--
Adrian Klaver
adrian.klaver@aklaver.com



--
--- Get your facts first, then you can distort them as you please.--

В списке pgsql-general по дате отправления:

Предыдущее
От: Brian Mills
Дата:
Сообщение: Re: [GENERAL] Recovery Assistance
Следующее
От: "David G. Johnston"
Дата:
Сообщение: [GENERAL] using hstore to store documents