100% cpu with wrong date cell

[eaglexiang]

We are handling excel files uploaded by customers every day, I mean those excel files are created by customers on net.

Lately we found wrong date-style cell may cause 100% cpu issue.

steps to reproduce:

1. create a blank excel file
2. set one cell as text format
3. enter a large number like someone's chinese id-card number 500123199005132245
4. set the cell as date format
5. read it with Rows.Columns() of excelize

Program will hang up because of countless time.Add, id-card number(500123199005132245) will be handled as days count.

// Duration is limited to aprox. 290 years
for intPart > MDD {
    durationDays := time.Duration(MDD) * time.Hour * 24
    date = date.Add(durationDays)
    intPart = intPart - MDD
}

I found a block about date in ECMA-376.

For compatibility, a SpreadsheetML application can interpret serial-number values in cells or in formulas
as dates. This subclause describes how serial number values can be converted to date values depending
on the compatibility mode.
A date that can be interpreted as a numeric value is a serial value. This is made up of a signed integer
date component and an unsigned fractional time component. Going forward in time, the date
component of a serial value increases by 1 each day. A serial value represents a UTC date and time, and,
as such, has no timezone information.

It may be right in ECMA-376 but danger in production environments. Stream Reader(Rows) is needed considering large file size and we cannot find a way to avoid the performance issue.

[eaglexiang]

time.Time holds seconds since Jan 1 year 1885 with uint64, whose maximum value is 18446744073709551615.

It means the max days count is 18446744073709551615/60/60/24 == 213503982334601.

obviously:

213503982334601 (since 1885)
<
500123199005132245 (since 1904 or 1899)

In another word, unlimited time.Add will cause int64 overflow.

If excelize can check it before loop-time.Time.Add, this issue would be solved :)

[xuri]

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

[eaglexiang]

Thanks for your work.

[imom0]

@eaglexiang 这些日期格式的身份证，是不是WPS来的？我们在业务里也遇到了，很怀疑是WPS做了什么“优化”。

就算这样解决了CPU占用的问题，也还是没拿到身份证原文吧。

[eaglexiang]

这些日期格式的身份证，是不是WPS来的？我们在业务里也遇到了，很怀疑是WPS做了什么“优化”。

我们处理的文件来自客户，并不清楚他们使用什么工具编辑这些文件。不过我在这个 issue 里有说复现方式。

就算这样解决了CPU占用的问题，也还是没拿到身份证原文吧。

现在是可以拿到的。

[imom0]

这些日期格式的身份证，是不是WPS来的？我们在业务里也遇到了，很怀疑是WPS做了什么“优化”。

我们处理的文件来自客户，并不清楚他们使用什么工具编辑这些文件。不过我在这个 issue 里有说复现方式。

就算这样解决了CPU占用的问题，也还是没拿到身份证原文吧。

现在是可以拿到的。

以 412822199406130000 为例解析出来 -258738860607/12/26 了

[eaglexiang]

这些日期格式的身份证，是不是WPS来的？我们在业务里也遇到了，很怀疑是WPS做了什么“优化”。

我们处理的文件来自客户，并不清楚他们使用什么工具编辑这些文件。不过我在这个 issue 里有说复现方式。

就算这样解决了CPU占用的问题，也还是没拿到身份证原文吧。

现在是可以拿到的。

以 412822199406130000 为例解析出来 -258738860607/12/26 了

刚刚验证了一下，你是对的，的确存在这个问题。但是这个恐怕无法通过 excelize 来解决，除非 excelize 能提供一个绕过格式解析，直接获取单元格原始值的方法 @xuri

[xuri]

The excel will display ######## if apply date or time number format for the cell which contains Chinese 18 digits identity card number, set the valid number format by SetCellStyle before reading the cell value by Excelize, for example re-set number format for A1:

f, err := excelize.OpenFile("Book1.xlsx")
if err != nil {
    fmt.Println(err)
    return
}
styleID, err := f.NewStyle(&excelize.Style{NumFmt: 1})
if err != nil {
    fmt.Println(err)
}
if err := f.SetCellStyle("Sheet1", "A1", "A1", styleID); err != nil {
    fmt.Println(err)
}