Delete sheet breaks AutoFilter

[pandaadb]

Description

When using autoFilter, removing Sheet1 seems to be breaking the entire file.

This is tested with LibreOffice on ubuntu.

When uploaded to Office365, it says the file has been repaired.

Steps to reproduce the issue:

Code:

        f := excelize.NewFile()

	f.NewSheet("sheet_test")
	for i := 0; i < 5; i++ {
		idx, _ := excelize.CoordinatesToCellName(i+1, 1)
		f.SetCellValue("sheet_test", idx, fmt.Sprintf("Header %d", i))
	}

	row := 2

	for j := 0; j < 10; j++ {

		for i := 0; i < 5; i++ {
			idx, _ := excelize.CoordinatesToCellName(i+1, row)
			f.SetCellValue("sheet_test", idx, fmt.Sprintf("Value %d", i))
		}

		row++
	}

	f.AutoFilter("sheet_test", "A1", "E11", "")

	f.DeleteSheet("Sheet1")
	f.SetActiveSheet(0)

	filename := "/tmp/test4.xlsx"

	f.SaveAs(filename)

Describe the results you received:

When opening the file, Libre office reports:

General input/output error

Describe the results you expected:

A 1 sheet xlsx file, containing the Autofilter.

Output of go version:

go version go1.13.10 linux/amd64

Excelize version or commit ID:

github.com/360EntSecGroup-Skylar/excelize/v2 v2.3.2

Environment details (OS, Microsoft Excel™ version, physical, etc.):

Ubuntu 18.04
LibreOffice (LibreOffice 6.0.7.3 00m0(Build:3)

[xuri]

Thanks for your issue. As the docs of the DeleteSheet noticed, this issue caused by some references not removed after deleting the worksheet which has the auto filter, I'll consider resolving it.

[pandaadb]

Oh that makes sense. I am sorry, I hadn't read the docs on DeleteSheet!

A question, is there a "safe" way to create a new spreadsheet without the first Sheet? I resolved to removing it because I had no idea how to get rid off it :)

[xuri]

If you needn't the default worksheet named Sheet1, you can rename it by SetSheetName("Sheet1", "sheet_test") before coming up with operations.

[xuri]

I have fixed it, please try to use the master branch code, and this patch will be released in the next version.