Thread: Docs claim that "select myTable.*" wildcard won't let you assign column names

<div style="font-family: Verdana;font-size: 12.0px;"><div> </div><div>From
http://www.postgresql.org/docs/9.4/static/sql-select.html(and previous version too):</div><div><h3><tt
class="COMMAND"> ##SELECT</tt> List</h3><div>    <...></div></div><div>    Instead of an expression, <tt
class="LITERAL">*</tt>can be written in the output list as a shorthand for all the columns of the selected
rows.</div><div>   Also, you can write <tt class="LITERAL"><tt class="REPLACEABLE c2">table_name</tt>.*</tt> as a
shorthandfor the columns coming from just that table. In these cases it is not</div><div>    possible to specify new
nameswith <tt class="LITERAL">AS</tt>; the output column names will be the same as the table columns'
names.</div><div> </div><div>But, thedocs elsewhere feature a query example show the use of a wildcard for
columns</div><div>aswell as allowing you to assign names to as many of the leading columns as you wish:</div><div> 
<div><div>WITHT0 as ( SELECT 1,2,3 )<br /> SELECT T0.* from T0 as T0(foo,bar) ;</div><div> </div><div><div> foo │ bar │
?column?<br/> ─────┼─────┼──────────<br />    1 │   2 │        3<br /> (1 row)</div><div> </div><div>The following
curiousvariant also works:</div><div> </div><div>WITH T0 as ( SELECT 1,2,3 )<br /> SELECT justAnythingReally.* from T0
asjustAnythingReally(foo,bar) ;</div></div>   <div>The synoposis/grammer at the top doesn't hint at this either. I've
checkedand this has been supported since at least 9.2
.</div><div> </div><div>Regards,</div><div>Amir</div><div> </div></div></div></div>

On Monday, September 21, 2015, Amir Rohan <amir.rohan@mail.com> wrote:

 

From http://www.postgresql.org/docs/9.4/static/sql-select.html (and previous version too):

  ##SELECT List

    <...>

    Instead of an expression, * can be written in the output list as a shorthand for all the columns of the selected rows.

    Also, you can write table_name.* as a shorthand for the columns coming from just that table. In these cases it is not

    possible to specify new names with AS; the output column names will be the same as the table columns' names.

 

But, the docs elsewhere feature a query example show the use of a wildcard for columns

as well as allowing you to assign names to as many of the leading columns as you wish:

 

WITH T0 as ( SELECT 1,2,3 )
SELECT T0.* from T0 as T0(foo,bar) ;

 

 foo │ bar │ ?column?
─────┼─────┼──────────
   1 │   2 │        3
(1 row)

 

The following curious variant also works:

 

WITH T0 as ( SELECT 1,2,3 )
SELECT justAnythingReally.* from T0 as justAnythingReally(foo,bar) ;

 

The synoposis/grammer at the top doesn't hint at this either. I've checked and this has been supported since at least 9.2 .

 

On Monday, September 21, 2015, Amir Rohan <amir.rohan@mail.com> wrote:


>> From http://www.postgresql.org/docs/9.4/static/sql-select.html[http://www.postgresql.org/docs/9.4/static/sql->>
select.html](and previous version too): 
>>
>>   ##SELECT List
>>
>>     <...>
>>     Instead of an expression, * can be written in the output list as a shorthand for all the columns of the
>>     selected rows.
>>     Also, you can write table_name.* as a shorthand for the columns coming from just that table. In these
>>     cases it is not possible to specify new names with AS; the output column names will be the same as the
>>     table columns' names.
>>
>> But, the docs elsewhere feature a query example show the use of a wildcard for columns
>> as well as allowing you to assign names to as many of the leading columns as you wish:
>>
>>
>> WITH T0 as ( SELECT 1,2,3 )
>> SELECT T0.* from T0 as T0(foo,bar) ;<...>

On Monday, September 21, 2015, David G. Johnston wrote:

> Neither of those examples is:
>
> SELECT * AS "how would one alias this?" FROM table
>
> So what's your point?

My point is that "In these cases it is not possible to specify new names with AS" is misleading because
it *is* possible and useful, but requires syntax which isn't clearly shown (if at all) where I'd expect it.
I think that could be improved.

> Obviously you can alias stuff before it makes its way into a select-list that refers to it using *

"obvious" to whom? probably not to someone who's level of SQL mastery has brought him/her to reading
the exciting "SELECT" documentation. I do see your point though, in the grammar the "AS" in my example
belongs not to `output_name ` but to `from_item`. So this syntax is hidden away behind the `column_alias`
production.

> In this case the FROM clause is what is being aliased.  It is documented though I'd need to look to
> identify the specific location.

This belongs in the page describing SELECT, and though I've looked I haven't found it. If I'm wrong (
I did look again just now), please correct me.

Amir

On 09/21/2015 06:04 PM, Amir Rohan wrote:
> This belongs in the page describing SELECT, and though I've looked I haven't found it. If I'm wrong (
> I did look again just now), please correct me.

Feel free to submit a doc patch if you feel it needs to be improved, but
it is mentioned here:

 http://www.postgresql.org/docs/9.4/interactive/sql-select.html#SQL-FROM

8<-----------
alias

    A substitute name for the FROM item [...] If an alias is written, a
column alias list can also be written to provide substitute names for
one or more columns of the table.
8<-----------

Joe

--
Crunchy Data - http://crunchydata.com
PostgreSQL Support for Secure Enterprises
Consulting, Training, & Open Source Development

On Monday, September 21, 2015, Amir Rohan <amir.rohan@mail.com> wrote:

On Monday, September 21, 2015, Amir Rohan <amir.rohan@mail.com> wrote:


>> From http://www.postgresql.org/docs/9.4/static/sql-select.html[http://www.postgresql.org/docs/9.4/static/sql->> select.html] (and previous version too):
>>
>>   ##SELECT List
>>
>>     <...>
>>     Instead of an expression, * can be written in the output list as a shorthand for all the columns of the
>>     selected rows.
>>     Also, you can write table_name.* as a shorthand for the columns coming from just that table. In these
>>     cases it is not possible to specify new names with AS; the output column names will be the same as the
>>     table columns' names.
>>
>> But, the docs elsewhere feature a query example show the use of a wildcard for columns
>> as well as allowing you to assign names to as many of the leading columns as you wish:
>>
>>
>> WITH T0 as ( SELECT 1,2,3 )
>> SELECT T0.* from T0 as T0(foo,bar) ;<...>

On Monday, September 21, 2015, David G. Johnston wrote:

> Neither of those examples is:
>
> SELECT * AS "how would one alias this?" FROM table
>
> So what's your point?

My point is that "In these cases it is not possible to specify new names with AS" is misleading because
it *is* possible and useful, but requires syntax which isn't clearly shown (if at all) where I'd expect it.
I think that could be improved.

> Obviously you can alias stuff before it makes its way into a select-list that refers to it using *

"obvious" to whom? probably not to someone who's level of SQL mastery has brought him/her to reading
the exciting "SELECT" documentation. I do see your point though, in the grammar the "AS" in my example
belongs not to `output_name ` but to `from_item`. So this syntax is hidden away behind the `column_alias`
production.

> In this case the FROM clause is what is being aliased.  It is documented though I'd need to look to
> identify the specific location.

This belongs in the page describing SELECT, and though I've looked I haven't found it. If I'm wrong (
I did look again just now), please correct me.

The docs don't make a big deal of alising other than noting individually where they can happen (cte, from clause, select-list) and few questions that are raised are probably better served to be simply handled on these lists than cluttering the docs.

On Monday, September 21, 2015, David G. Johnston wrote:

>To be honest, if you have a situation where you think you need to alias "*"
>you should probably be expanding that "*" out into individual column names
>anyway, which you can easily alias.
>

That depends. If my table has 30 columns and I just need to rename
the first couple, then I probably would find that advice unreasonable.

>The docs don't make a big deal of alising other than noting individually
>where they can happen (cte, from clause, select-list) and few questions
>that are raised are probably better served to be simply handled on these
>lists than cluttering the docs.

What constitutes "clutter", or what is "obvious", really  depends on who is
making the assertion. A pg veteran's views probably don't mirrort those of the
person you'd typically find actually reading the docs, trying to figure out how
to write the query they want.

But yeah,  the pg documentation isn't the best resource to pick up SQL from
(nor is is their purpose), and no patch I'd submit would really change that.
So, we're done.

Amir