FOR XML Issue With Numbers Coming Across In Scientific Format
Dec 5, 2006
Program Descritpion:
Our product allows our customers to enter their product catalogs into our ASP.NET pages and we save the data using SQL Server 2005. Customer use GridViews to edit their data and teh data is broken into manageable groupings (called Field Groups) stored as meta-data in SQL Server Tables. Then when a user wants to edit a set of data they choose the Field Group and we dynamically generate a data source and a gridview, bind them together, and our users update their data. A bit more complicated but that's teh gist.
Another thing they can do, and where our problem occurs, is they can print reports or do exports into excel using this data. Since there are so many fields, they use the Field Groups to select which fields they want included so everything needs to be built dynamcially, at runtime, using the Field Group meta-data. Essentially, I want my grid to have these 10 fields.
Problem:
For the export we generate an XML data source using FOR XML in SQL Sprocs. The xml is pulled into a xmldatadocument, trasnformed with an xsl file into an Excel XML Spreadsheet. Our problem is that FOR XML is generating our XML real numbers in scientific format. The xsl "format-number" function does not recognize scientific format and returns NaN (not a number) instead of the value in my spreadsheet. If I leave it blank I get the scientific number but Excel doesn't format it correctly, the cell has an error tag that wants me to choose Number stored as text or convert to a number. I need it to show up on teh Excel form already formatted without that message.
I can't change the DataType of the field in SQL, too many other things depend on it and it needs to be a real. I can't use CONVERT(decimal,fieldName) in SQL because the SQL string is dynmically generated using Dynamic SQL and most of the fields are not real. When we build the Field List we just have field names, we can't check anything to add CONVERT functions to only real fields.
Is there any way to force the XML output to not be scientific for the entire document? Or another function in XSL I'm unaware of (pretty new to xsl)? Or perhaps something I can do in the XML Spreadsheet tags to force the conversion?
I have a table which stores date-of-birth in varchar 19861231(yyyymmdd). A view takes this data. I want to store this date as mmddyyyy in the view. How can we achieve this?
SELECT membername, outputval case when choice = 0 then outputval else null end as outputval from MyDatabase group by membername, outputval
how to format outputval: if outputval < 40000 format outputval as: 5 - 5.78 - 6.9 - 6,778 - 4,567.8 - 12,456.78 - etc. if outputval >= 40000 format it as a scientific.
Hello Nice simple T-SQL question that will no doubt prove totally uncontroversial :) Anyone know the easiest way to get from:123456789to123,456,789in T-SQL? Currently I cast my int as money, use convert to varchar with style 1 and then lop off the .00. Is there an easier way? Two things before you get all upset and start squealing about front ends:1) These are admin functions that I am running from SSMS. I don't want to go to the bother of creating an application when SSMS is perfectly adequate for what I need (namely to-a-large-degree unformatted result sets).2) I'm not too fussed about it just curious. Ta & tra la!
Hi all i'm having an issues when i trying to do a report:
i make a report and there is a field which contains data that can be numeric and numeric with letter: "77756" or "345WS" when i export my report to excel i and i open it i get like a make in the upper left coner of the cells that have only numbers asking me to convert this to numeric format, but this think is that i need this values to be string so i can apply filtering. is there a way that i can format my cell to string so the excel take it like that and allowme to do the autoifilter.
To be able to tell the Period number ([PD]) to change from an integer to a var_char so that it can be concatenated with the Fiscal Year. I.e. Fiscal Year || Period number to produce the following results.
2006 01
2006 02
2006 03
Etc……
Status of Problem:
The issue is that I can only get it done to the point to where the Fiscal Year || Period number yields the following results:
20061
200610
200611
200612
20062
Etc….
This is as issue since I am trying to get all the information from the past to a certain year and period. Using a filter that says
Fiscal Year + Period Number <= ?Prompt_Fiscal_Year? + ?Prompt_Period_Number?
So if I enter Year: 2007 Period 3
I should get:
2006 01
2006 02
2006 03
Etc…….
2006 11
2006 12
2007 01
2007 02
2007 03
Can you please let me know if my thinking is off or if my SQL is wrong, Thank you for your assistance in this matter.
I would like to know if there is any trick to convert numbers to csv files. I don't have any control to format the numbers.
Source: Sql server
for example:
select cast( 1 as float)/ cast(201 as float)
Destination:CSV file
the output of the file is 4,9751243E-3
i don't want to have scientific notation on the number. I want the result to be 0,00497512437810945. how can i do it?
I tried to change the type of destination columns to String or numeric but nothing works. what is the influence of changing data type in the destination columns e the output to files?
I tried using convert but with the same results.
Can it be because of local settings or something else?
/* Format strings for fractional Numbers and Currency values
Return Values: VARCHAR Parameters: @n Specifies numeric expression to format. @sFormat: Specifies one or more format codes that determine how the expression is formatted.
The following table lists the available format codes.
Hello, I'm having problems with a column of numeric string data coming from a tab delimited CSV file. When a number in the column is negative, it is expressed this way; 1,240.52-
The negative symbol occurs at the end of the numeric string. The destination column is in a SQL Server 2005 table. I've tried to change the data type of the table column, and three or four different data types in a data conversion task, but nothing has worked so far to bring the data over.
I have table with 3 columns: record_id (int), car (varchar(50)), amount (decimal(18,4)).
I created dimension for drillthrough action. Also I set Amount's column FormatString to '##0.00' but when I use Drillthough column Amount values is shown like:
.0000 15.7900 18.4100
Is there any possibility to change that values would be shown as follow:
I have an excel export with numeric values. When the SSIS package writes into the excel it treats all data types as strings and attaches an apostrophe in the beginning. I tried formatting the excel sheet with the number data type and saving it . It doesnt work.
Other strange thing is that when I go into the advanced editor for the excel destination and look into the properties of the external columns all of them have the data type as Unicode String[DT_WSTR] irrespective of what the data type is from the input.The UI allows me to change it to numeric for numeric columns without any error but the value is not saved at all which is very frustrating. I also tried changing it in the xml file of the package, it some how seems to overwrite it after validation. It would be really nice if it threw an error saying that it cannot be changed.
Anyways there seems to be no way of changing the destination data type if it s an existing file and if I create a new excel sheet there is no way of formatting data. Is there any way out?
I have a report with a column which contains either a string such as "N/A" or a number such as 12. A user exports the report to Excel. In Excel the numbers are formatted as text.
I already tried to set the value as CDbl which returns error for the cells containing a string.
The requirement is to export the column to Excel with the numbers formatted as numbers and the strings such as "N/A' in the same column as string.
I've got some values stored in nvachar(255) field stored by mistake as scientific notation (eg 7.5013e+006 instead of 7501301) and I need to convert and update the field with normal entry, not scientific notation. Is there a way to do that?
Hi, I am new to SQL 2005. I have to design schema for scientific data warehouse. Data is available in 2 or more flat data files recorded at 1 sec interval. At Least 2 of the data files have 100+ columns. I am inclined to create a table per data file type. I want to know If this is correct/optimal for me to do?
I don't think I can create normalize tables based on the headers in these Data files.
Primary Objective of this data warehouse is make it available for reporting services and Analysis Services.
HiWe've got some numbers stored as Reals which are returning values inscientific notation that we need rounded down to 3 digits to the rightof the decimal.ie 8.7499999E-2 needs to return 8.75.Round, cast, convert, formatnumber in the dts package all fail.Help!Thanks Moe
Hi, I am new to SQL 2005. I have to design schema for scientific data warehouse. Data is available in 2 or more flat data files recorded at 1 sec interval. At Least 2 of the data files have 100+ columns. I am inclined to create a table per data file type. I want to know If this is correct/optimal for me to do?
I don't think I can create normalize tables based on the headers in these Data files.
Primary Objective of this data warehouse is make it available for reporting services and Analysis Services.
We sometimes have small values stored in a column with datatype of float like 0.000644470739403048 which is being converted to -5.8E-05. Perhaps that is OK to be stored in the database however I need the value in decimal format to use. (I'm using longitude values in google maps).
is there anything I can do at the database level. I was looking at the properties which is 53 numeric precision and 8 length.
I'm trying to find a way to format a FLOAT variable into a varchar inSQL Server 2000 but using CAST/CONVERT I can only get scientificnotation i.e. 1e+006 instead of 1000000 which isn't really what Iwanted.Preferably the varchar would display the number to 2 decimal placesbut I'd settle for integers only as this conversion isn't businesscritical and is a nice to have for background information.Casting to MONEY or NUMERIC before converting to a varchar works finefor most cases but of course runs the risk of arithmetic overflow ifthe FLOAT value is too precise for MONEY/NUMERIC to handle. If anyoneknows of an easy way to test whether overflow will occur and thereforeto know not to convert it then that would be an option.I appreciate SQL Server isn't great at formatting and it would be fareasier in the client code but code this is being performed as adescription of a very simple calculation in a trigger, all stored tothe database on the server side so there's no opportunity for clientintervention.Example code:declare @testFloat floatselect @testFloat = 1000000.12select convert(varchar(100),@testFloat) -- gives 1e+006select cast(@testFloat as varchar(100)) -- gives 1e+006select convert(varchar(100),cast(@testFloat as money)) -- gives1000000.12select @testFloat = 12345678905345633453453624453453524.123select convert(varchar(100),cast(@testFloat as money)) -- givesarithmetic overflow errorselect convert(varchar(100),cast(@testFloat as numeric)) -- givesarithmetic overflow errorAny suggestions welcome...CheersDave
I'm developing a scientific appliation - to forecast the spread of disease. The algorithm has been designed such that different subpopulations reside on separate computers.
The model uses the same set of data repeatedly, and only returns a small signature that reflects the appropriateness of a particular set of parameters.
The model does however change, and I'm looking for a way to roll out this model (which only contains T-SQL, CLR assemblies and service broker code).
I have considered setting up a database on each computer which has the 'data' files which are static, and a separate database which contains the algorithm which can be amended.
Could one then detach the algorithm database making the new version that can then be rolled out (by attaching it to each computer used)?
Is this the best practice? Does anyone have a suggestion as to how to do this?
Am I missing something or is there something odd with float data types. i know that float is not the most precise definition but i came across something really odd today.
first let me define the scenario.
this is sql server 2005, standard edition build 3042.
I have a table defined as
CREATE TABLE [dbo].[ASSET](
[Property_Num] [numeric](10, 0) NOT NULL,
[Accrual_Factor_Val] [float] NULL
)
the accrual_factor_val was updated to a value of 0.00005 then the web service failed because the proc returned 5E-05!
i opened the table, and discovered this is the stored value. is this correct?
Why does M$ Query Analyzer display all numbers as positive, no matterwhether they are truly positive or negative ?I am having to cast each column to varchar to find out if there areany negative numbers being hidden from me :(I tried checking Tools/Options/Connections/Use Regional Settings bothon and off, stopping and restarting M$ Query Analyer in betwixt, butno improvement.Am I missing some other option somewhere ?
I have a table with a column ID of ContentID. The ID in that column is all NULLs. I need a way to change those nulls to a number. It does not matter what type of number it is as long as they are different. Can someone point me somewhere with a piece of T-SQL that I could use to do that. There are over 24000 rows so cursor change will not be very efficient.
I have an 'ID' column. I'm up to about ID number 40000, but not all are in use, so ID 4354 might not be in any row. I want a list of all numbers which aren't in use. I want to write something like this:
select [numbers from 0 to 40000] where <number> not in (select distinct id from mytable)
Coming from Sybase env. I would like to know the equivalent of sp_showplan. What it does is, it shows the showplan for the query being run by spid, which is passed as a parameter to sp_showplan.
I've got an employee table with a date of birth field in it. i need a query that will a allow me to list all employees who's birthdays are coming up the next 30 days (or 1 month, if easier). I've tried several approaches & am getting nowhere... Any help would be greatly appreciated.
I just read the release notes for SQL 2005 SP2 CTP and the 'Select All' on multi-select parameters is coming back EXACTLY as it was in the first release. If this is not the case then I apologise for the rant and request that someone clarify the situation but otherwise .... I had to change loads of reports because it was removed in SP1, now I'll have to change them again for SP2.
Please Microsoft, stop doing this to us!
At least make it a check box option somewhere. The reasons for it's removal in SP1 still exist, but it is a nice feature to have, so just make it configurable.
I will probably have to delay installing SP2 at various customer sites because it will make my parameter lists look stupid. This is a shame because there are some useful updates in it.
