Often we get data which is distributed in multiple column sets. Here is how you can convert it to good tabular data using Power Query. Watch the video. Sample File download.
This is the worst type of input data to get. Each row has been split into two (or sometimes more) rows. We want those to be combined into a single row. In an earlier article, I have shown one method of doing this.
Now here is a simpler, faster and more powerful method using Power Query.
(Estimated reading time 12 min)
Download the Presentation (3 MB). I have added lots of links to related articles and external references including books written by stalwarts in Excel analytics field.
Thank you for the overwhelming response. More than 800 attendees, 70+ questions.
Watch the video on YouTube (use 720P) or Download (MP4, 188 mb, 60 min) and view on your PC. A big thank you to the Microsoft and Economic Times team for making this possible.
Power Query is an amazing tool for importing and cleaning up data from various sources. The problem was that it required you to have professional version of Office. Now that problem is solved…
Two simple approaches. One is to add an extra column with serial numbers and then sort on that column in descending order. This works for small amount of data. For large data it is best to import it in Power Query and choose Transform tab – Reverse Rows.
Benefit of Power Query? It works on a smaller sample of data and then applies the transformation when you choose Save and Load option. This is much faster than getting all the data and then trying to sort it (which is the first method).
Why is this required? Usually required with logs where the first transactions or rows are at the bottom. So the data is received in reverse chronological order. Twitter feeds, Timeline Updates, Live blogs – all follow this pattern.
This method works independent of the time-stamp column. What is wrong with timestamp? It may be in different time formats, some rows may have same timestamp and some rows may have no timestamp at all.