Missing Record - Phantom Record

[Andrew]

Hi All,

Have come across something weird and am after some help.

Say i run this query where rec_id is a column of table arlhrl,

select * from arlhrl where rec_id >= 14260

This returns to me 2 records with rec_id's of 14260 and 14261

Then I run this query

select * from arlhrl where rec_id >= 14263

This returns 7 records with rec_ids of 14263 up.

How come the first query doesn't return the records returned by the
2nd query also?

If I select for 14262 no records are returned. It is like this is a
phantom record or has an end of file character in it.

I tried re-creating the indexes but to no avail. If anyone has any
ideas about what could be causing it or how to fix it it would be much
appreciated.

Thanks in advance,

Andrew

[Andrey]

Andrew wrote:

Hi All,

Have come across something weird and am after some help.

Say i run this query where rec_id is a column of table arlhrl,

select * from arlhrl where rec_id >= 14260

This returns to me 2 records with rec_id's of 14260 and 14261

Then I run this query

select * from arlhrl where rec_id >= 14263

This returns 7 records with rec_ids of 14263 up.

How come the first query doesn't return the records returned by the
2nd query also?

If I select for 14262 no records are returned. It is like this is a
phantom record or has an end of file character in it.

I tried re-creating the indexes but to no avail. If anyone has any
ideas about what could be causing it or how to fix it it would be much
appreciated.

Thanks in advance,

Andrew

Hi,

First, stupid question - is the field 'rec_id' of integer type?
Why i am asking is because i had a similar example myself when i started with my new job - i was
quering an id field and got weird results as you do. Then i found that some 'smart ass' made this
comlumn a varchar for no reason - just because she was doing like that in Access all the time before :)
Second, what i'd do when i get into an unexplainable glitch:

SELECT * INTO <new table> FROM <your table>

And try to query the records from the new table without setting any indexes - just as is - as you
know SELECT INTO just copies raw data without any underlying stuff.
See what you'll get.
From my experience there are a of of people who are allowed to mess with SQL databases but don't
have a clue what they are doing, and when you start using their 'smart ideas' sometimes it's just
hard to follow their logic :) So maybe some setting were set a wrong way somewhere, you can never
imagine what another person could do - believe me, i just got quite a few awsome examples within the
last month since i got this job :)

Let me know how it works!

Andrey

[Andrew]

Andrey <le*******@yaho o.com> wrote in message news:<7wt3d.787 69$D%.11878@att bi_s51>...

Andrew wrote:

Hi All,

Have come across something weird and am after some help.

Say i run this query where rec_id is a column of table arlhrl,

select * from arlhrl where rec_id >= 14260

This returns to me 2 records with rec_id's of 14260 and 14261

Then I run this query

select * from arlhrl where rec_id >= 14263

This returns 7 records with rec_ids of 14263 up.

How come the first query doesn't return the records returned by the
2nd query also?

If I select for 14262 no records are returned. It is like this is a
phantom record or has an end of file character in it.

I tried re-creating the indexes but to no avail. If anyone has any
ideas about what could be causing it or how to fix it it would be much
appreciated.

Thanks in advance,

Andrew

Hi,

First, stupid question - is the field 'rec_id' of integer type?
Why i am asking is because i had a similar example myself when i started with my new job - i was
quering an id field and got weird results as you do. Then i found that some 'smart ass' made this
comlumn a varchar for no reason - just because she was doing like that in Access all the time before :)
Second, what i'd do when i get into an unexplainable glitch:

SELECT * INTO <new table> FROM <your table>

And try to query the records from the new table without setting any indexes - just as is - as you
know SELECT INTO just copies raw data without any underlying stuff.
See what you'll get.
From my experience there are a of of people who are allowed to mess with SQL databases but don't
have a clue what they are doing, and when you start using their 'smart ideas' sometimes it's just
hard to follow their logic :) So maybe some setting were set a wrong way somewhere, you can never
imagine what another person could do - believe me, i just got quite a few awsome examples within the
last month since i got this job :)

Let me know how it works!

Andrey

Hi Andrey,

Thanks for your reply. I tried as you mentioned, inserting into new
table etc but to no avail. I did figure out what the problem was
though.

This particular table had been upsized from a foxpro table. One of the
columns in the foxpro table had a maximum value of numeric 9999.
Somehow, someone had tried to insert a value large than this so foxpro
put in ****. On the upsize, and I can only assume here, sql must have
thought 'hang on, you must mean infinity here' and put a bit-wise
pattern (1.#INF) for infinity into this particular column for the
record.

This only became evident when using Enterprise Manager and returning
all rows on the given table, it did display the record with the value
1.#INF in the column for the 'missing' record. As to why it displayed
in EM and not Query Analyser is anyone's guess, but surely the queries
that led me to this initial discovery shouldn't have behaved like
this!!??

posting

http://groups.google.com/groups?q=%2...gle.com&rnum=1

gives some ideas.

Thanks anyway,

Andrew

[Andrey]

Andrew wrote:

Andrey <le*******@yaho o.com> wrote in message news:<7wt3d.787 69$D%.11878@att bi_s51>...

Andrew wrote:

Hi All,

Have come across something weird and am after some help.

Say i run this query where rec_id is a column of table arlhrl,

select * from arlhrl where rec_id >= 14260

This returns to me 2 records with rec_id's of 14260 and 14261

Then I run this query

select * from arlhrl where rec_id >= 14263

This returns 7 records with rec_ids of 14263 up.

How come the first query doesn't return the records returned by the
2nd query also?

If I select for 14262 no records are returned. It is like this is a
phantom record or has an end of file character in it.

I tried re-creating the indexes but to no avail. If anyone has any
ideas about what could be causing it or how to fix it it would be much
appreciate d.

Thanks in advance,

Andrew

Hi,

First, stupid question - is the field 'rec_id' of integer type?
Why i am asking is because i had a similar example myself when i started with my new job - i was
quering an id field and got weird results as you do. Then i found that some 'smart ass' made this
comlumn a varchar for no reason - just because she was doing like that in Access all the time before :)
Second, what i'd do when i get into an unexplainable glitch:

SELECT * INTO <new table> FROM <your table>

And try to query the records from the new table without setting any indexes - just as is - as you
know SELECT INTO just copies raw data without any underlying stuff.
See what you'll get.
From my experience there are a of of people who are allowed to mess with SQL databases but don't
have a clue what they are doing, and when you start using their 'smart ideas' sometimes it's just
hard to follow their logic :) So maybe some setting were set a wrong way somewhere, you can never
imagine what another person could do - believe me, i just got quite a few awsome examples within the
last month since i got this job :)

Let me know how it works!

Andrey

Hi Andrey,

Thanks for your reply. I tried as you mentioned, inserting into new
table etc but to no avail. I did figure out what the problem was
though.

This particular table had been upsized from a foxpro table. One of the
columns in the foxpro table had a maximum value of numeric 9999.
Somehow, someone had tried to insert a value large than this so foxpro
put in ****. On the upsize, and I can only assume here, sql must have
thought 'hang on, you must mean infinity here' and put a bit-wise
pattern (1.#INF) for infinity into this particular column for the
record.

This only became evident when using Enterprise Manager and returning
all rows on the given table, it did display the record with the value
1.#INF in the column for the 'missing' record. As to why it displayed
in EM and not Query Analyser is anyone's guess, but surely the queries
that led me to this initial discovery shouldn't have behaved like
this!!??

posting

http://groups.google.com/groups?q=%2...gle.com&rnum=1

gives some ideas.

Thanks anyway,

Andrew

Well, EM and QA might show you diferent results because they are using diferent methods of 'talking'
to sql server.
QA is using isql.com, precisely it's isqlw.com version, which is an old DB lib based way of connection.
EM, i guess, is using ODBC or OLEDB connection.
I also had a headache not long time ago, when i used sql console tools to make Python work with sql
server. I had a table with varcha fields which had around couple thousand characters of text each.

When i used isql.com to retreive those text records, text returned truncated, around 300 to 600
characters left.. SO i started using osql.com instead, and no headache.

So resume is - every time you're in doubt, use both EM and QA

PS. BTW, I didn't know sql server can store 'infinity' values. Thanks for the info!

WYGL,
Andrey

[Andrey]

Andrew wrote:

Andrey <le*******@yaho o.com> wrote in message news:<7wt3d.787 69$D%.11878@att bi_s51>...

Andrew wrote:

Hi All,

Have come across something weird and am after some help.

Say i run this query where rec_id is a column of table arlhrl,

select * from arlhrl where rec_id >= 14260

This returns to me 2 records with rec_id's of 14260 and 14261

Then I run this query

select * from arlhrl where rec_id >= 14263

This returns 7 records with rec_ids of 14263 up.

How come the first query doesn't return the records returned by the
2nd query also?

If I select for 14262 no records are returned. It is like this is a
phantom record or has an end of file character in it.

I tried re-creating the indexes but to no avail. If anyone has any
ideas about what could be causing it or how to fix it it would be much
appreciate d.

Thanks in advance,

Andrew

Hi,

First, stupid question - is the field 'rec_id' of integer type?
Why i am asking is because i had a similar example myself when i started with my new job - i was
quering an id field and got weird results as you do. Then i found that some 'smart ass' made this
comlumn a varchar for no reason - just because she was doing like that in Access all the time before :)
Second, what i'd do when i get into an unexplainable glitch:

SELECT * INTO <new table> FROM <your table>

And try to query the records from the new table without setting any indexes - just as is - as you
know SELECT INTO just copies raw data without any underlying stuff.
See what you'll get.
From my experience there are a of of people who are allowed to mess with SQL databases but don't
have a clue what they are doing, and when you start using their 'smart ideas' sometimes it's just
hard to follow their logic :) So maybe some setting were set a wrong way somewhere, you can never
imagine what another person could do - believe me, i just got quite a few awsome examples within the
last month since i got this job :)

Let me know how it works!

Andrey

Hi Andrey,

Thanks for your reply. I tried as you mentioned, inserting into new
table etc but to no avail. I did figure out what the problem was
though.

This particular table had been upsized from a foxpro table. One of the
columns in the foxpro table had a maximum value of numeric 9999.
Somehow, someone had tried to insert a value large than this so foxpro
put in ****. On the upsize, and I can only assume here, sql must have
thought 'hang on, you must mean infinity here' and put a bit-wise
pattern (1.#INF) for infinity into this particular column for the
record.

This only became evident when using Enterprise Manager and returning
all rows on the given table, it did display the record with the value
1.#INF in the column for the 'missing' record. As to why it displayed
in EM and not Query Analyser is anyone's guess, but surely the queries
that led me to this initial discovery shouldn't have behaved like
this!!??

posting

http://groups.google.com/groups?q=%2...gle.com&rnum=1

gives some ideas.

Thanks anyway,

Andrew

And how did you get rid of that infinity value in the in field?

[Andrew]

I got rid of the infinity value using EM open table then typed in the
value I wanted.

Andrey <le*******@yaho o.com> wrote in message news:<pir4d.287 70$wV.19066@att bi_s54>...

Andrew wrote:

Andrey <le*******@yaho o.com> wrote in message news:<7wt3d.787 69$D%.11878@att bi_s51>...

Andrew wrote:

Hi All,

Have come across something weird and am after some help.

Say i run this query where rec_id is a column of table arlhrl,

select * from arlhrl where rec_id >= 14260

This returns to me 2 records with rec_id's of 14260 and 14261

Then I run this query

select * from arlhrl where rec_id >= 14263

This returns 7 records with rec_ids of 14263 up.

How come the first query doesn't return the records returned by the
2nd query also?

If I select for 14262 no records are returned. It is like this is a
phantom record or has an end of file character in it.

I tried re-creating the indexes but to no avail. If anyone has any
ideas about what could be causing it or how to fix it it would be much
appreciate d.

Thanks in advance,

Andrew
Hi,

First, stupid question - is the field 'rec_id' of integer type?
Why i am asking is because i had a similar example myself when i started with my new job - i was
quering an id field and got weird results as you do. Then i found that some 'smart ass' made this
comlumn a varchar for no reason - just because she was doing like that in Access all the time before :)
Second, what i'd do when i get into an unexplainable glitch:

SELECT * INTO <new table> FROM <your table>

And try to query the records from the new table without setting any indexes - just as is - as you
know SELECT INTO just copies raw data without any underlying stuff.
See what you'll get.
From my experience there are a of of people who are allowed to mess with SQL databases but don't
have a clue what they are doing, and when you start using their 'smart ideas' sometimes it's just
hard to follow their logic :) So maybe some setting were set a wrong way somewhere, you can never
imagine what another person could do - believe me, i just got quite a few awsome examples within the
last month since i got this job :)

Let me know how it works!

Andrey

Hi Andrey,

Thanks for your reply. I tried as you mentioned, inserting into new
table etc but to no avail. I did figure out what the problem was
though.

This particular table had been upsized from a foxpro table. One of the
columns in the foxpro table had a maximum value of numeric 9999.
Somehow, someone had tried to insert a value large than this so foxpro
put in ****. On the upsize, and I can only assume here, sql must have
thought 'hang on, you must mean infinity here' and put a bit-wise
pattern (1.#INF) for infinity into this particular column for the
record.

This only became evident when using Enterprise Manager and returning
all rows on the given table, it did display the record with the value
1.#INF in the column for the 'missing' record. As to why it displayed
in EM and not Query Analyser is anyone's guess, but surely the queries
that led me to this initial discovery shouldn't have behaved like
this!!??

posting

http://groups.google.com/groups?q=%2...gle.com&rnum=1

gives some ideas.

Thanks anyway,

Andrew

And how did you get rid of that infinity value in the in field?