Forum Discussion
Import Report output file
Hi
I have snapshot report output files stored in a folder. Files in the folder:
ReportID_100 12012016.txt
...
ReportID_100 11012019.txt
ReportID_100 12012019.txt
ReportID_100 01012020.txt
...
Each file is like as below:
ReportID_100 Report Name: "Antibiotics Group" Report Date: 12/01/2019 Report Page 1
11/01/2019 - 11/30/2019
Prescriptions Group by Antibiotics by Department
Prescription PSCP 2 Count PSCP 3 Count PSCP 4 and above Count
---------------------------------------------------------------------------------
Radiology 1 0 1
Outpatient 1 0 1
Pathology 0 0 0
Surgery 1 0 1
Medicine 1 0 1
Orthopedics 1 0 1
ReportID_100 Report Name: "Antibiotics Group" Report Date: 12/06/2019 Report Page 2
11/01/2019 - 11/30/2019
Prescriptions Group by Antibiotics by Department
Prescription PSCP 2 Count PSCP 3 Count PSCP 4 and above Count
---------------------------------------------------------------------------------
Adult/Family Medicine 1 0 1
Dermatology 1 0 1
Gastroenterology 1 0 1
...
ReportID_100 Report Name: "Antibiotics Group" Report Date: 12/06/2019 Report Page 3
11/01/2019 - 11/30/2019
Prescriptions Group by Antibiotics by Department
Prescription PSCP 2 Count PSCP 3 Count PSCP 4 and above Count
---------------------------------------------------------------------------------
Urology 1 0 1
Pulmonary Medicine 1 0 1
...
Total: 11 10 222
#END OF REPORT#
I am having multiple pages data, page numbers (dynamic) and date range in each file. I am trying to bring and model the data.
a) Any tips on modeling of importing this kind of files?
b)I am able to get the data from each file and parse. But, how to identify the header rows in between pages and strip them of? (as each file has first few rows dynamic and also comes in between pages)
ReportID_100 Report Name: "Antibiotics Group" Report Date: 12/06/2019 Report Page 2
11/01/2019 - 11/30/2019
Prescriptions Group by Antibiotics by Department
Prescription PSCP 2 Count PSCP 3 Count PSCP 4 and above Count
---------------------------------------------------------------------------------
Thanks in advance
3 Replies
Hi sevenhills ,
I'm unable to go into too much detail around your specific scenario as it will require some quite creative transformations, I think. However, your starting point should be to use either the Folder or SharePointFolder data sources in Power Query. Once you've selected your folder with all these reports in it, you will get the option to Combine & Transform them:
This will allow you to select an example file (ideally one that includes all the issues you'd like to fix across the group), and perform transformations on it that will be automatically replicated across all the files in that folder, before they are all finally appended together into a single query.
While I can't specifically advise on the actual transformations you would need to do, my guess would be that you would use basic filters on certain words to remove header rows, use 'Fill Up' and 'Fill Down' functions to move values up and down to rows that aren't going to be filtered out, and functions like Text.Contains etc. to identify certain words/phrases to be retained.
Again, I think you're going to have to get pretty creative to knock these into a workable shape, but this is the process you need to complete what you require.
Pete
Thanks for your reply.
My question was related to and adding more details:
a) Any tips on modeling of importing this kind of files?
More details:
- I have 100,000 files for over 150 reports, in one linux folder. I got them to Azure.
- I created power bi data flows, one per report type and contains multiple queries in each dataflow
- Note: we can have all in one data flow instead of one per report type. But decided to go one flow per report type.
b)I am able to get the data from each file and parse. But, how to identify the header rows in between pages and strip them of? (as each file has first few rows dynamic and also comes in between pages)
More details:
- Before posting and asking in the forum, I am able to do the combine and transform files per report.
- Each snapshot report output file has header rows in between pages
- Issue is dynamicness involved in each file to process the header rows in each page. I want to strip them.
- After stripping all the header and footer rows, it will be easy create data as columns and use for visualization.
Hope this clarifies the needs 🙂
Thanks
For (b), using power query, I solved it as below
* Step 1: Analyze and Identified the sets as (hard coded for each report)
sets: page header, table header, footer row, page end row ...
This is more for categorizing
* Step 2:
Get data all the files, and line number in each file
File Name Row Number File 1 1 File 1 2 File 1 3 File 1 … File 2 1 File 2 2 File 2 3 File 2 … File 3 1 File 3 2 File 3 3 File 3 … For each file, go through each row and based on the data identified first row of each set
* Step 3: Remove the dynamic text involved in these rows like removing page number. say,
ReportID_100 Report Name: "Antibiotics Group" Report Date: 12/06/2019 Report Page 2as
ReportID_100 Report Name: "Antibiotics Group" Report Date: 12/06/2019 Report Page* Step 4: look the whole table with these category rows and mark the first row (and last row of each set) and create summary of rows for each file as that need to be deleted as
Delete Row Details as
File Name Set Name Row Begin Row End File 1 Page Header 1 6 File 1 Table Header 7 9 File 1 Page Header 22 27 File 1 Table Header 28 31 ... File 1 Report End 1011 1011 * Expand by adding custom column to the table:
{[Row Begin]..[Row End]}* Later did left anti join to get only data rows.
🙂Hope it helps for others later
(moving this code logic to dataflow, which is outside this post, FYI)
