1

Topic: SQL Developer: wrong sorting of the Roman numbers with a point (VI.I)

Version SQL Developer
Versioning 4.2.0.17.089.1709
Build 17.089.1709
Clearly, as should be, and there is a suspicion that it is possible to adjust numerical or regional settings on request runtime.

select V. I ' as n from dual union all
select ' VI.I I ' as n from dual union all
select ' VI.II I ' as n from dual union all
select VII. n'as n from dual union all
select VIII. I ' as n from dual
order by n
----------The answer
V. I
VI.I I
VII. I
VI.II I
VIII. I

And in real request (there in sorting this column of the second) - and there 7 generally earlier 6
In  all , and in it is more convenient, but at me now to the Developer access is easier.
It would not be desirable to drag order_index in finite , what it is possible to adjust?

2

Re: SQL Developer: wrong sorting of the Roman numbers with a point (VI.I)

exp98;
In oracle there is no built in type the Roman numbers, therefore there is a sorting as lines
The sorting order of characters of lines is regulated by parameter
NLS_SORT, which if explicitly it is not installed, depends on adjustment NLS_TERRITORY
toad uses Oracle Client adjustments, SQL*Developer adjustments java which pulls these  from system.

wrote:

And in real request (there in sorting this column of the second) - and there 7 generally earlier 6
In  all , and in it is more convenient, but at me now to the Developer access is easier.

I do not trust, show result,  here it is used banal ASCII7, and the sorting difference should not be

3

Re: SQL Developer: wrong sorting of the Roman numbers with a point (VI.I)

Only it seems to you that it is numbers. And the assumption about "" the Roman number as produces lines possession of the account at level of the first class.

4

Re: SQL Developer: wrong sorting of the Roman numbers with a point (VI.I)

exp98;
Try     to the wished

5

Re: SQL Developer: wrong sorting of the Roman numbers with a point (VI.I)

[spoiler NLSSORT c union all without inline view does not work?]

column n format a30
with s as (
select v. I ' as n from dual union all
select ' vi.i I ' as n from dual union all
select ' vi.ii I ' as n from dual union all
select vii. n'as n from dual union all
select viii. I ' as n from dual)
select *
from s
order by nlssort (n, ' NLS_SORT=BINARY ');
N
------------------------------
v. I
vi.i I
vi.ii I
vii. I
viii. I
select v. I ' as n from dual union all
select ' vi.i I ' as n from dual union all
select ' vi.ii I ' as n from dual union all
select vii. n'as n from dual union all
select viii. I ' as n from dual
order by n;
N
------------------------------
v. I
vi.i I
vii. I
vi.ii I
viii. I
select v. I ' as n from dual union all
select ' vi.i I ' as n from dual union all
select ' vi.ii I ' as n from dual union all
select vii. n'as n from dual union all
select viii. I ' as n from dual
order by nlssort (n, ' NLS_SORT=BINARY ');
order by nlssort (n, ' NLS_SORT=BINARY ')
*
Error in line 6:
ORA-01785: argument ORDER BY should be number of expression of the SELECT-list

[/spoiler]

6

Re: SQL Developer: wrong sorting of the Roman numbers with a point (VI.I)

Vadim Lejnin wrote:

in oracle there is no built in type the Roman numbers

DocID 236825.1

7

Re: SQL Developer: wrong sorting of the Roman numbers with a point (VI.I)

andrey_anonymous;
Cool! Did not know, thanks

8

Re: SQL Developer: wrong sorting of the Roman numbers with a point (VI.I)

AmKad wrote:

NLSSORT c union all without inline view does not work?

Yes, too ORA-01785.
At the Developer not only a point. As it is exact it and digit instead of a point, and ' _ ' instead of it too. I already after a post checked up it.
All the same thanks I will look at variants, calms that on working basis through the Client will work (Vadim Lejnin). If that I will tighten order_index.
Vadim Lejnin, do not trust? See a picture
This time villages Elik in a pool (too after all tells that villages in the machine, instead of in a pool...)

9

Re: SQL Developer: wrong sorting of the Roman numbers with a point (VI.I)

Concerning link DocID 236825.1
Please, teach, as it to transform into the link for the browser.

10

Re: SQL Developer: wrong sorting of the Roman numbers with a point (VI.I)

exp98 wrote:

Concerning link DocID 236825.1
Please, teach, as it to transform into the link for the browser.

1. To get at responsible in your organization access on support.oracle.com
2. To hammer the specified identifier of the document into a line of search
Unfortunately, the publication of these documents here is forbidden as , and forum rules.

11

Re: SQL Developer: wrong sorting of the Roman numbers with a point (VI.I)

exp98 wrote:

At the Developer not only a point. As it is exact it and digit instead of a point, and ' _ ' instead of it too. I already after a post checked up it.
All the same thanks I will look at variants, calms that on working basis through the Client will work (Vadim Lejnin). If that I will tighten order_index.
Vadim Lejnin, do not trust? See a picture
This time villages Elik in a pool (too after all tells that villages in the machine, instead of in a pool...)

1. NLSSORT works differently depending on a NLS-environment. Yes, the punctuation in some surroundings all is collected in a heap.
Also is not present, it not property of the client or application - is simple they at you work with a different NLS-environment.
More in detail - in Globalization Support Guide
2. Also what on a picture?
Suppose, what VI really should follow for VII?
VII
VI
VIII
3. And here about Elic it is vain - it, unlike you, subject NLS knows thoroughly.

12

Re: SQL Developer: wrong sorting of the Roman numbers with a point (VI.I)

I see.

13

Re: SQL Developer: wrong sorting of the Roman numbers with a point (VI.I)

andrey_anonymous;
I assumed another that sorting does not throw out characters. And with abortions all is logical.
About Elika I am right, if it was not thin  from its side. I hope, from it does not decrease.

14

Re: SQL Developer: wrong sorting of the Roman numbers with a point (VI.I)

exp98 wrote:

About Elika I am right, if it was not thin  from its side.

Well try to refute it, here to you a data set:

with t as (select rownum N, to_char (r, ' RN ') RN from dual connect by level <1000)
select * from t
order by rn

15

Re: SQL Developer: wrong sorting of the Roman numbers with a point (VI.I)

andrey_anonymous;
I was not going to develop a subject of the Maestro... I know that did not read  particulars about sorting, I much that did not read and at all I do not suspect about . Your example to refute probably it does not turn out.
Thanks all for clarification and time found for me.
All flew in a head instantly then I and carried out that verdict more low, it is possible not to read.
[spoiler] to suppose to dispute the end, consider a phrase literally:
"Only it seems to you that it is numbers. And the assumption about"  "the Roman number as produces lines possession of the account at level of the first class."
1 sentence read literally, incorrectly, it not seemed to me instants.
2 sentence in the most consistent interpretation:
Sending I assume that it is a question the Roman numbers. Incorrectly, see above.
Sending speech about number (?) sorting as lines, and I so assumed. From this the output about not possession of the account, though Arabian, though Roman, though is for some reason drawn... r-adicheskim.  I did not assume that it is a question of numbers, though what. Again, as you understood, an incorrect output. That I wrote about numerical adjustments, I meant especially a point as a separator for normal decimal notation, and that suddenly it somehow affects.
And generally and here any account?
Sending
Unique rational grain in 2nd sentence, with .. my ignorance of a subject , could consist that , meeting in the field VI.I, shows "", and that this behavior is connected with NLS. Also that this behavior meant between lines. But also in this case, the output about possession of the account extremely  is connected to a hint. And besides, and here the account?
.. The phrase as a whole is "a manure heap", to dig in a k-plenty it is possible only at desire by all means to find pearl grain and at confidence that it there was valid. But so far as the Maestro  meant it and  so estimated my knowledge here a question, in what period it estimated duration of my digging?
Starting with enough big for a small phrase kol-va unsuccessful sendings, I instantly decided, it is not necessary what to search for depth sense (and already and it is not necessary), and mine  the remark in comparison with presumable "assumptions" of the Maestro of me, looks quite innocently and an obvious joke. Try to refute.
Whether and it was necessary to search for depth sense, the author knows only, but to calculate for truth it is improbable, and in cultural community it is better not to continue this subject.
[/spoiler]

16

Re: SQL Developer: wrong sorting of the Roman numbers with a point (VI.I)

Vadim Lejnin wrote:

andrey_anonymous;
Cool! Did not know, thanks

What for numbers?
Knew  month and "year" (rm/rn)
ps
access on a metalink for me closed
....
stax

17

Re: SQL Developer: wrong sorting of the Roman numbers with a point (VI.I)

Stax wrote:

  access on a metalink for me closed

I and here to you closed access.

18

Re: SQL Developer: wrong sorting of the Roman numbers with a point (VI.I)

  wrote:

it is passed...
I and here to you closed access.

To  all goes, not long remains
.....
stax

19

Re: SQL Developer: wrong sorting of the Roman numbers with a point (VI.I)

Stax wrote:

it is passed...
What for numbers?
Knew  month and "year" (rm/rn)
ps
access on a metalink for me closed
....
stax

And there just about a year also goes
Your message of thirteen-year prescription operation with dates
Well and inverse function it is sensitive another

20

Re: SQL Developer: wrong sorting of the Roman numbers with a point (VI.I)

Ljubomudrov wrote:

And there just about a year also goes
Your message of thirteen-year prescription operation with dates
Well and inverse function it is sensitive another

Thanks
But  not absolutely built in type
.....
stax