xpath //*[.="Foo"] 和 xpath //*["Foo"] 在谓词中只有一个字符串有什么区别？

Question

Doing some testing in iPython with a new (to me) way of selecting nodes based on their text in an xpath. (Irrelevant lines omitted for clarity)

In [26]: from lxml import etree

In [41]: string = '''
    ...: <outer>
    ...:    <mid>
    ...:       <inner>Foo</inner>
    ...:    </mid>
    ...: </outer>
    ...: '''

In [43]: root = etree.fromstring(string)

In [44]: root.xpath('//inner[text()="Foo"]')
Out[44]: [<Element inner at 0x10a0387c0>]

In [45]: root.xpath('//inner[.="Foo"]')
Out[45]: [<Element inner at 0x10a0387c0>]

In [47]: root.xpath('//inner["Foo"]')
Out[47]: [<Element inner at 0x10a0387c0>]

That all makes sense to me so far. However:

In [48]: root.xpath('//*[text()="Foo"]')
Out[48]: [<Element inner at 0x10a0387c0>]

In [49]: root.xpath('//*[.="Foo"]')
Out[49]: [<Element inner at 0x10a0387c0>]

In [50]: root.xpath('//*["Foo"]')
Out[50]: 
[<Element outer at 0x10a188200>,
 <Element mid at 0x10a01d6c0>,
 <Element inner at 0x10a0387c0>]

I had expected the second and third xpaths to produce the same result by matching all three nodes. Can anyone explain why they're different?

Answer 1

Spec says

A PredicateExpr is evaluated by evaluating the Expr and converting the result to a boolean. If the result is a number, the result will be converted to true if the number is equal to the context position and will be converted to false otherwise; if the result is not a number, then the result will be converted as if by a call to the boolean function . Thus a location path para[3] is equivalent to para[position()=3] .

(emphasis mine). Thus

root.xpath('//*["Foo"]')

is equivalent to

root.xpath('//*["Lemon Pie"]')

It does not test for content of your <inner> node; in fact, because "Foo" is a truthy literal, AFAIK it is also equivalent to

root.xpath('//*')

As Barmar said, the first and second expression do not match nodes other than <inner> because of whitespace. To get all three, trim (or "normalize space", in XPath language):

root.xpath('//*[normalize-space()="Foo"]')

Answer 2

To add to Amadan's answer, your first and second expressions are not equivalent. Both of them will match

<Inner>Foo</Inner>

but they give different results for constructs such as

<Inner><span>Foo</span></Inner>

or

<Inner>Foo<nbsp/>Bar</Inner>

As a general rule, 9 times out of 10 when someone uses text() they should change it to . .