I have this set of data that contains hyphens sign in my cell. My plan is to replace this hyphens sign to a number of 0 so that I can convert my column to numeric data type to do a calculation. The tricky part is where I only need to replace the cell that only contains hyphens sign (row 2 and row 3) without changing the negative value (row 6,7,8 and 9). I tried to use
regexp_replace("M1KW",'-','0','g')
function but it does change the negative value as well.
Any other special command that I can use ?
This is my sample data:
You seem to want to treat the string '-' as '0'. If so, you can directly replace the values:
update t
set mikw = '0'
where mikw = '-';
so that I can convert my column to numeric data type to do a calculation
You don't need to UPDATE all rows before being able to convert the data type to the correct one:
alter table the_table
alter mikw type numeric
using (case when mikw = '-' then 0 else mikw::numeric end);
That will do the "update" and the change of the data type in a single operation, rather than two.
Seems like the RIGHT function would be useful here, for isolating only instances where the hyphen is the last character.
Try this out:
UPDATE your_table
SET M1KW = 0
WHERE RIGHT(M1KW, 1) = '-';
Related
Example of the data in csv
Column_header
000000025000{
000000007185E
The doucmention I have
*The right most position of the numeric field will have a sign OVER it
designating positive or negative.
Example of Data
I dont understand how write the logic to support the the symbol,number,letter to get the correct value.
I'd create a table (or view) with the static mapping of character-value, meaning:
Symbol
Value
J
-1
A
+1
about the data rows themselves, it seems to me there is always a symbol at the end, therefore you can split the data into two columns, value, and symbol...
I have no idea about how the data are inserted but it seems logically easy
SELECT
_YourValue_
,LEFT(_YourValue_, LENGTH(_YourValue_)-1) as Value
,RIGHT(_YourValue_, 1) as Symbol
FROM _Whatever_
you can also cast to whatever datatype is correct for those data.
Finally you can join the tables and show/calculate whatever is needed
select value , if(value LIKE '%{%' or value LIKE '%J%' or value LIKE '%E%' or value LIKE '%C%',concat(SUBSTRING(value,1,char_length(value)-1),'+'),concat(SUBSTRING(value,1,char_length(value)-1),'-')) as new_value from yourtablename
Output
value
New Value
000000025000{
000000025000+
000000007185E
000000007185+
Add all other character on first parameter of if clause for positive designation.
CASE
WHEN code IN ('FJS354', 'JDF334')
THEN 'Lower_form'
ELSE 0
END AS format
This returns an error in Redshift
invalid input syntax for integer: "Lower_form"
I know if I change 'Lower_form' to an integer it will work however I want this column to be a string. Is there a way to do this?
I want this column to be a string.
All branches of a case expression must return the same datatype. You are giving two literal values whose datatype is not the same (string vs integer): the database makes the decision to turn them both to integers - which is not what you want.
Rremove the ambiguity by being explicit about the datatype you want to return. That is, make this literal 0 a string:
CASE WHEN code in ('FJS354','JDF334')
THEN 'Lower_form'
ELSE '0'
END as format
I have a specific column in a table, it shall contains only numbers in Nvarchar that have a length of 3. Unfortunately, some users wrote '12' but they should have written '012'. There were not enough validation at the time.
I need to fix that. Here is the logic I used :
UPDATE [Mandats_Approvisionnement].[dbo].[ARTICLE_ECOLE]
SET [UNIT_ADM] = STUFF(UNIT_ADM, 0, 0, '0')
WHERE LEN(UNIT_ADM) = 2;
The error goes like :
Cannot insert the value NULL into column 'UNIT_ADM', table
'Mandats_Approvisionnement.dbo.ARTICLE_ECOLE'; column does not allow
nulls. UPDATE fails.
I can't see where the problem is, I verified and all the records contain at least 2 characters, so the STUFF function cannot returns null as there are no NULL records in that table column [unit_adm]... How do I make it work ?
It should be stuff(UNIT_ADM,1,0,'0') as stuff returns null if the start position is 0.
Citing the documentation:
If the start position or the length is negative, or if the starting
position is larger than length of the first string, a null string is
returned. If the start position is 0, a null value is returned.
You could make this simpler by using
right('0' + UNIT_ADM, 3)
instead of stuff.
I have a product table called PDPRODP - for certain styles within this table I used a concat statement to add a full-stop to their description (PRDESC), I now wish to remove this full stop.
The descriptions are varying length, the field max size is 30 characters and I need to physically remove the full-stop rather than using a select statement to trim the full-stop.
I tried;
UPDATE PDPRODP SET PRDESC = PRDESC-1 where PRSTYLE = 1234
But I got this error:
Character in CAST argument not valid.
I also tried this following some googling;
UPDATE PDPRODP SET PRDESC=LEFT(PRDESC, LEN(PRDESC)-1)
WHERE PRCOMP = 1 AND PRSTYL = 31285
But got this error:
LEN in *LIBL type *N not found.
Use LENGTH
UPDATE PDPRODP SET PRDESC=LEFT(PRDESC, LENGTH(PRDESC)-1)
WHERE PRCOMP = 1 AND PRSTYL = 31285
The REPLACE() function can search for all occurrences of some string, and substitute another in its place. You might search for your full-stop, and replace it with a zero-length string ''. This would be handy in cases where your search string may not always be at the end.
I import Excel files via SSIS to SQL-Server. I have a temp table to get everything in nvarchar. For four columns I then cast the string to money type and put in my target table.
In my temp table one of those four columns let me call it X has a comma as the delimiter the rest has a dot. Don't ask me why, I have everything in my SSIS set the same.
In my Excel the delimiter is a comma as well.
So now in my target table I have everything in comma values but the X column now moves the comma two places to the right and looks like this:
537013,00 instead of 5370,13 which was the original cell value in the temp and excel column.
I was thinking this is a culture setup problem but then again it should or shouldn't work on all of these columns.
a) Why do I receive dot values in my temp table when my Excel displays comma?
b) how can I fix this? Can I replace the "," in the temp table with a dot?
UPDATE
I think I found the reason but not the solution:
In this X column in excel the first three cells are empty - the other three columns all start with 0. If I fill these three cells of X with 0s then I also get the dot in my temp table and the right value in my target table. But of course I have to use the Excel file as is.
Any ideas on that?
Try the code below. It checks whether the string value being converted to money is of numeric data type. If the string value is of numeric data type, then convert it to money data type, otherwise, return a NULL value. And it also replaces the decimal symbol and the digit grouping symbol of the string value to match the expected decimal symbol and digit grouping symbol of SQL Server.
DECLARE #MoneyString VARCHAR(20)
SET #MoneyString = '$ 1.000,00'
SET #MoneyString = REPLACE(REPLACE(#MoneyString, '.', ''), ',', '.')
SELECT CAST(CASE WHEN ISNUMERIC(#MoneyString) = 1
THEN #MoneyString
ELSE NULL END AS MONEY)
As for the reason why you get comma instead dot I have no clue. My first guess would be cultural settings but you already checked that. What about googling, did you get some results?
First the "separator" in SQL is the decimal point: its only excel that is using the comma. You can change the formatting in excel: you should format the excel column as money and specify a decimal point as the separator. Then in the SSIS import wizard split out the transformation of the column so it imports to a money data type. Its a culture thing, but delimiter tends to be used in the context of signifying the end of one column and the start of the next (as in csv)
HTH
Well thats a longstanding problem with excel. It uses the first 30 or so rows to infer data type. It can lead to endless issues. I think your solution has to be to process everything as a string in the way Yaroslav suggested, or supply an excel template to have data predefined and formatted data type columns, which then have the values inserted. Its a pita.