Our challenge is the following one :
in an Azure SQL database, we have multiple tables with the following table names : table_num where num is just an integer. These tables are created dynamically so the number of tables can vary. (from table_1, table_2 to table_N) All tables have the same columns.
As part of a U-SQL script file, we would like to execute the same query on all of these tables and generate an output csv file with the combined results of all these queries.
We tried several things :
U-SQL does not allow looping so we were thinking creating a View in our Azure SQL database that would combine all the tables using a cursor of some sort. Then, the U-SQL file would query this View (using external source). However, a View in Azure SQL database can only be created via a function and a function cannot execute dynamic SQL or even call a stored procedure...
We did not find a way to call a stored procedure of the external data source directly from U-SQL
we dont want to update our U-SQL job each time a new table is added...
Is there a way to do that in U-SQL through a custom extractor for instance? Any other ideas?
One solution I can think of is to use Azure Data Factory (v2) to assist in this.
You could create a pipeline with the following activities:
Lookup activity configured to execute the stored procedure
For Each activity that uses the output of the lookup activity as a source
As a child item use a U-Sql Activity that executes your U-Sql script which writes the output of a single table (the item of the For Each activity) to blob or datalake
Add a Copy Activity that merges the blobs from step 2.1 to one final blob.
If you have little or no experience working with ADF v2 do mind that it takes some time to get to know it but once you do, you won't regret it. Having a GUI to create the pipeline is a nice bonus.
Edit: as #wBob mentions another (far easier) solution is to somehow create a single table with all rows since all dynamically generated table have the same schema. You can create a stored procedure for populating this table for example.
Related
I am trying to create a dataflow in SSIS where the source data originates from an excel file and reaches to a temporary staging table in a SQL server where I can add various stored procedures to the data.
The dataflow that I have created stores the data permanently on what is supposed to be the staging area.
I would like to get some ideas on creating the staging table in SQL with the SSIS dataflow.
your question is a bit confusing. I suppose that you are maybe trying to make the data loaded in the table of the staging area temporary without keeping the past loaded data.
If I'm right what you're trying to accomplish is a "full resfresh" data flow.
From your description I assume you alerady have the staging table (so no nedd to CREATE it) but you need to truncate it at every run. You can achive this by using a Execut SQL Task element to the control flow with a TRUNCATE TABLE <YOUR TABLE NAME> in it. The data flow loading the data must be in dependency of this task with the result of truncating your table at every run.
If you need to CREATE a table you can do it in the control flow with the Execute SQL Task (you can execute any kind of query with this task), rember to set correctly the connection manager of the task.
We curate data in the "Dev" Azure SQL Database and then currently use RedGate's Data Compare tool to push up to 6 higher Azure SQL Databases. I am trying to migrate that manual process to ADFv2 and would like to avoid copy/pasting the 10+ copy data actives for each database (x6) to keep it more maintainable for future changes. The static tables have some customization in the copy data activity but the basic idea follows this post to perform an upsert.
How can the implementation described above be done in Azure Data Factory?
I was imagining something like the following:
Using one parameterized link service that has the server name & database name configurable to generate a dynamic connection to Azure SQL Database.
Creating a pipeline for each table's copy data activity.
Creating a master pipeline to then nest each table's pipeline in.
Using variables loop over the different connections an passing those to the sub-pipelines parameters.
Not sure if that is the most efficient plan or even works yet. Other ideas/suggestions?
we can not tell you if that's the most efficient plan. But I think so. Just make it works.
As you said in the comment:
we can use Dynamic Pipelines - Copy multiple tables in Bulk with
'Lookup' & 'ForEach'. we can perform dynamic copies of your data
table lists in bulk within a single pipeline. Lookup returns either
the lists of data or first row of data. ForEach - #activity('Azure
SQL Table lists').output.value ;
#concat(item().TABLE_SCHEMA,'.',item().TABLE_NAME,'.csv') + This is
efficient and cost optimized since we are using less number of
activities and datasets.
In usually, we also will choose same solution with you: dynamic parameter/pipeline, lookup + foreach active to achieve the scenario. In one word, make the pipeline has a strong logic, simple and efficient.
Added the same info mentioned in the Comment as Answer.
Yup, we can use Dynamic Pipelines - Copy multiple tables in Bulk with 'Lookup' & 'ForEach'.
We can perform dynamic copies of your data table lists in bulk within a single pipeline. Lookup returns either the lists of data or first row of data.
ForEach - #activity('Azure SQL Table lists').output.value ;
#concat(item().TABLE_SCHEMA,'.',item().TABLE_NAME,'.csv')
This is efficient and cost optimized since we are using less number of activities and datasets.
Attached pic as ref-
I have a scenario to copy output of GET Metadata activity into a SQL table. Can I do this directly without using Databricks notebook?
You can make use of look up activity.
GetMetadata -> Lookup
And write insert SQL statement in Query, or use stored procedure.
I am writing an ETL to extract data from HANA table and load into SQL Server in BODS.
My job is to create a new table on SQL Server every time I run my job with name as date of that day. I know we can do that for flat files by using global variable but not sure how we can declare similar variable in template table to get desired results?
Why you want to use template tables. You can do the same as below:
Load the data in a standard staging table using BODS
Using DS scripting mechanism generate a query to create a table
Execute the query using SQL transform
Generate another query to copy data from staging table to the table created above
Several other ways also like you can write a DB procedure to create a table with the desired name and copy over the data from stage to that table. This procedure you can call from DS.
Hope this helps.
Cheers.
Shaz
In one SQL Task can I create a table variable
DELCARE #TableVar TABLE (...)
Then in another SQL Task or DataSource destination and select or insert into the table variable?
The other option I have considered is using a Temp Table.
CREATE TABLE #TempTable (...)
I would prefer to use Table Variable so that it remains in memory. But can use temp table if it is not possible to use table variable. Also I cannot use the record set destination as I need to preform straight SQL tasks on it later on.
The use case that this is trying to solve is essentially performing a transformation in the stead of BizTalk. There is a very large flat file to flat file transformation that BizTalk has to transform unfortunately the data volume would produce unacceptable load on the BizTalk server so the idea is to off load it to SSIS. However, it is not a simple row to row transformation, there are different types of rows which have relations to each other. The first task in SSIS is to load the row into appropriate (temp) tables, then in the second data task a select is preformed with the correct format for output.
You could use some of the techniques in this post: http://consultingblogs.emc.com/jamiethomson/archive/2006/11/19/SSIS_3A00_-Using-temporary-tables.aspx
especially the ones about using RetainSameConnection=TRUE on the connection manager.
I would be interested to see more information about what use case you have that requires you to write out data to a temp table or table variable before further SSIS processing. Couldn't you take care of all of the SQL required steps in your source query before you start processing the dataflow with SSIS?
Table variables are not kept solely in memory and can be written to disk under memory pressure. I tend to use table variables for very small lookups. If you Need to push a table into SQL Server due to necessary and complex transformations, then use a 'permanent' temp table that is truncated within the SSIS package prior to insert. Simple and will get what you need done.
The SSIS package would be run in a job. I assume it runs inside a SQL job. In that case, using a temp table won't harm. SQL Jobs are generally run after office hours so it does not matter.