Series articles ：

I often listen to others Python How powerful in the data field , As a result, I studied for a long time , Even data processing is a death of trouble . Only later , It's not Python Data processing is powerful , But he has a data analysis artifact —— pandas

Preface

Many never touched pandas After seeing the relevant information , I think this thing can't handle, like Excel Data in that complex and changeable format . Actually pandas Compared with manual operation 、 PivotTable operations are even Vba operation , It can process complex shape data more flexibly and quickly .

This time, , So let's see how it works pandas Generate various format data such as payslip at one time .

---

Case study

Continue to use the transcript data ：

We want to list each student's transcript separately , That is, a row of records becomes a small table ：

---

Excel skill

use Excel Can the built-in function complete this work in batch ？ Certainly. , It just takes a lot of skill . But I always say , Ideas are more important than methods , This is because , If you understand this Excel The operation process , The same idea can be applied to pandas Finish in .

have a look Excel How to accomplish this requirement , We use it 10 A record to demonstrate ：

• Add auxiliary sequence , Each value is separated by 2( This is because each row of the resulting data is separated by 2 That's ok )

---

• Under the auxiliary column , Generating sequences ( The end value is the mantissa of the previous sequence )
• Use Excel The built-in generation series function can

---

• Batch add header row
• from 3 Start , every other 3 Set a title for the row

---

• Remove duplicate sequence lines

---

• Finally, sort according to the auxiliary column , You get the result

---

pandas practice

pandas There are many ways to implement , This section begins with the above Excel The solution is basically the same way . as follows ：

• Modify row index value , amount to Excel The first step

---

• reindex You can give non-existent row index values , Then those lines will become empty
• amount to Excel The second and fourth steps in

---

• Last , From the index 3 Start , every other 3 Just assign row and column headings

---

The complete code is as follows ：

---

Last

The method introduced in this paper has great practical limitations , For example, it cannot be divided according to any dimension , And it requires a user pair pandas Have a deep understanding of the index in . Next time, we will introduce a general way , You can also add summary rows to each small table .

If you want to learn from scratch pandas , Then you can look at my pandas special column .