I have a informatica mapping as follows
The source is oracle database and at the source qualifier I'm doing simple sum of all columns using SQL; out of which one column has bigint datatype.
At the target is greenplum database that should store the outcome of sum.
The row where it exceeds 19 digit sum is ignored by informatica and entered in rejected rows. Is there a way to avoid this problem other than changing the datatype of bigint at source qualifier or at target?
Have a look at Transformation Data Types. Bigint allows precision of up to 19 digits. For larger ones I guess you need to use Decimal. Have you tried that?
Related
All documentations and tutorials out there mention that transforming a table column to Data Type 'Fixed Decimal' will make it have four digits to the right of the period symbol.
However, when I use this type, it only shows two and the Power Query for that step has:
= Table.TransformColumnTypes(myTable,{{"fieldName", Currency.Type}})
Was the default behaviour changed or do I need to first set some setting to get four? I tried this with multiple files.
I am taking data from the excel where there is a column "Global" with values as whole numbers and decimal numbers. When I am importing it to Power BI, the column data type is coming as "TEXT" and hence I am unable to use it to calculate my new column which is "Local + Global / Total" since it cannot convert calculate on string which makes complete sense.
I looked it up and tried the following :
To change the column type to Whole Number / Decimal Number and it showed me an error that It cannot be converted.
I made a measure using the functions CONVERT and VALUE but in vain. It could not be converted saying a single value was expected but multiple values were provided in the latter.
How do I solve this?
EDIT
The column looks like this
Global
12345.67
43566
123.765
If it is showing error that It cannot be converted, obviously there are some garbage value in the column like - space, string or other values not a number. Please check your data first, hope you will get the issue as well.
In Power Query Editor You can select the column and change data type to Whole Number. In that case, some errors will be shown where the conversion failed to convert some value as shown below-
Here I have provided a string in the last row. Now you can check your data by filtering the column with error to check what are the actual values in the source. You can than perform some transformation to get the correct value value out of them.
You can also remove all records with error as shown below if you find those rows are with garbage value is not important for your.
I am new to Power BI and I am making a matrix for Sales value. As we use various currencies and convert the numbers, each Sales number can have as many as 15 decimal points. NumericColumn corresponds to Sales going forward.
There is this OriginalTable:
NumericColumn,
OtherColumn1,
... ,
OtherColumnN
This OriginalTable is imported from SQL Server and has ~ 400K rows. NumericColumn has type numeric(35,15).
As there are a lot of unimportant columns in OriginalTable I made a new table in the following manner where M << N:
NewTable = SUMMARIZECOLUMNS(
OriginalTable[NumericColumn],
OriginalTable[OtherColumn1],
... ,
OriginalTable[OtherColumnM]
)
The sum of the "NumericColumn" in OriginalTable is 2,550,832.98 . This number is seen by putting "NumericColumn" in Value field as Sum and without setting any Rows or Columns in Visualization -> Matrix.
On the other hand, the sum of the "NumericColumn" in NewTable is 2,550,829.14 (3.84 smaller than that in OriginalTable).
It seems that I lost precision when I created a new table. Would somebody kindly take a look? Please let me know if more information is needed.
Thanks & Regards,
Kyoto
By default, the M query transformation keeps up to 8 decimal points when you are importing data and applying Change Type. As you said you have up to 15 decimal points, you can try the below steps and check it worked for you or not.
Come back to report clicking Cloase & Apply button. Now in the report, select your decimal value column configure the column Format and Decimal points as shown below-
Now applying SUMMARIZECOLUMNS should give your expected result.
I have been trying to generate a calculated column within Power BI Desktop using the MEDIAN formula to get the median age. I checked to see if there were any inconsistencies like text data or blank data within the Age column but, did not find any such instances. The error message is as follows -
Expressions that yield variant data-type cannot be used to define
calculated columns.
The data used in the analysis is hosted on data.world
Any help will be highly appreciated.
This is a pretty strange situation but I think I found the explanation on the Microsoft Community Forum.
When the underlying column is of data type Whole Number, MEDIAN function returns a variant data type because it may return a Whole Number when there is no interpolation or a Decimal Number when there is interpolation. While measures can be of variant data type, calculated columns must be of a single data type, hence the error. To force MEDIAN to always return Decimal Number, change the expression to MEDIANX(Table1, [Column2] * 1.0).
Another fix would be to convert the Age column to the decimal number type rather than the whole number type. I also found wrapping with VALUE work.
I've imported data (approximately 200 columns) into Power BI desktop (latest version as of 2017-08-02) and have explicitly told the app to treat a number of columns as being percentages. Within the query editor, I can verify that my values are treated as such:
When I put my data into a table, they show up as normal floats, not percentages. When I click on the exact same column as in the above picture and view it in the Modeling tab, Power BI shows it as being "General" format:
While I can go through and change the formatting here to have them all be percentages, I have already done so in the query editor! Is there a way to make PBI recognize my already specified format?
Short answer: No.
Explanation:
In the query editor, you didn't actually specify any format. What you specified is the data type, so that the source data can be read correctly. Say you have a column with data like 001, you can specify it as text type so you can retain the leading zero.
However, the actual formatting (i.e. data presentation) is done in your second step, because even if it's a (decimal) number, you can still format it as a percentage, with different decimal places, etc. (vice versa)