XML Viewer and Formatter

Continue in the tool

More pages for this tool

xpath tester online xml

Query XML with XPath: examples that work in the tool

On the Tree tab, switch the search to XPath, write the expression and click a result to jump straight to the node. The examples below use the three-book catalog that ships as the tool sample.

The slice of XPath 1.0 the tool understands

XPath describes a path through the tree. A leading slash starts at the document root (/catalogo/livro), two slashes search at any depth (//autor), the at sign picks attributes (//livro/@id), the asterisk matches any element and text() returns the text. Filters go in square brackets: a number picks by position ([2]), last() takes the last one, [@categoria="design"] compares an attribute, [preco>100] compares a child's value, and the functions contains, starts-with, not and count, together with and and or, combine conditions. Position is counted among the siblings that passed the name test, as in the standard.
Names without a prefix match the local name, so //Envelope finds soap:Envelope without you writing the prefix, and with a prefix the match is exact. This differs from a library's XPath, which requires registering namespaces, and was chosen for quick testing. Named axes (ancestor::, following-sibling::), union with a vertical bar, variables and most functions are out; using them gives a warning with the position of the offending piece. Quotes can be single or double, and the result comes in document order, without repeats, capped at 200 items in the list.
For a NF-e XML, //det/prod/xProd lists the product names, //ICMSTot/vNF returns the total and //det[prod/vProd>200] filters the items above 200. Because the result is highlighted in the tree and the node line is shown, you can test an expression here before taking it to a Python script, an XSLT or a spreadsheet. Evaluation has a work ceiling, and an expression too heavy for the file is stopped with a message instead of freezing the page.

Expressions tested on the sample catalog

All books

Input: /catalogo/livro
Expected output: 3 resultados

Absolute path from the root.

Filter by attribute

Input: //livro[@categoria='programacao']/titulo
Expected output: 2 resultados

Returns the titles of the algorithms and networks books.

Numeric comparison

Input: //livro[preco>100]/titulo
Expected output: Redes na Prática

Only the 120.00 book passes.

Last item

Input: //livro[last()]/@id
Expected output: b3

last() is the size of the sibling set.

A child value

Input: //livro[autor='Ana Souza']
Expected output: 2 resultados

Compares the text of the autor element.

The contains function

Input: //titulo[contains(.,'Web')]
Expected output: Tipografia na Web

The dot is the node itself, here the title text.

Full tool FAQ

How do I open an XML file?

Drop the file on the upload area or pick it with the button. The tool detects the encoding (UTF-8, UTF-16 or the one declared in the header, such as ISO-8859-1), shows the formatted text and builds the tree. You can also paste the content. Any text editor or browser can open XML, but only here do you get formatting, a tree, search and NF-e reading in one place with nothing to install.

What does malformed XML mean and how is the error pointed out?

What is the difference between well-formed and valid XML?

Is my file uploaded to a server?

How do I read the XML of a Brazilian electronic invoice (NF-e)?

Does the NF-e reading work for NFC-e, CT-e and MDF-e?

Does formatting or minifying change the XML content?

Is it safe to open XML from an unknown source here?

What is the maximum file size?

How do I use XPath in the search?

Frequently asked questions

Which XPath version does the tool use?

A subset of XPath 1.0: paths, //, attributes, positions, last(), comparisons, contains, starts-with, not, count, and and or. Named axes and union are not included.

How do I query an XML with a namespace?

Write just the local name, such as //numero, and it matches ex:numero. To restrict to one prefix, write the full name, //ex:numero.

Can I export the XPath result?

Click a result to select the node and use the buttons that copy the path and the value. The list shows each hit's full path, ready to paste.

Next steps

The slice of XPath 1.0 the tool understands

XPath describes a path through the tree. A leading slash starts at the document root (/catalogo/livro), two slashes search at any depth (//autor), the at sign picks attributes (//livro/@id), the asterisk matches any element and text() returns the text. Filters go in square brackets: a number picks by position ([2]), last() takes the last one, [@categoria="design"] compares an attribute, [preco>100] compares a child's value, and the functions contains, starts-with, not and count, together with and and or, combine conditions. Position is counted among the siblings that passed the name test, as in the standard.

Names without a prefix match the local name, so //Envelope finds soap:Envelope without you writing the prefix, and with a prefix the match is exact. This differs from a library's XPath, which requires registering namespaces, and was chosen for quick testing. Named axes (ancestor::, following-sibling::), union with a vertical bar, variables and most functions are out; using them gives a warning with the position of the offending piece. Quotes can be single or double, and the result comes in document order, without repeats, capped at 200 items in the list.

For a NF-e XML, //det/prod/xProd lists the product names, //ICMSTot/vNF returns the total and //det[prod/vProd>200] filters the items above 200. Because the result is highlighted in the tree and the node line is shown, you can test an expression here before taking it to a Python script, an XSLT or a spreadsheet. Evaluation has a work ceiling, and an expression too heavy for the file is stopped with a message instead of freezing the page.

Expressions tested on the sample catalog

All books

Input: /catalogo/livro
Expected output: 3 resultados

Absolute path from the root.

Filter by attribute

Input: //livro[@categoria='programacao']/titulo
Expected output: 2 resultados

Returns the titles of the algorithms and networks books.

Numeric comparison

Input: //livro[preco>100]/titulo
Expected output: Redes na Prática

Only the 120.00 book passes.

Last item

Input: //livro[last()]/@id
Expected output: b3

last() is the size of the sibling set.

A child value

Input: //livro[autor='Ana Souza']
Expected output: 2 resultados

Compares the text of the autor element.

The contains function

Input: //titulo[contains(.,'Web')]
Expected output: Tipografia na Web

The dot is the node itself, here the title text.

Full tool FAQ

How do I open an XML file?

What does malformed XML mean and how is the error pointed out?

What is the difference between well-formed and valid XML?

Is my file uploaded to a server?

How do I read the XML of a Brazilian electronic invoice (NF-e)?

Does the NF-e reading work for NFC-e, CT-e and MDF-e?

Does formatting or minifying change the XML content?

Is it safe to open XML from an unknown source here?

What is the maximum file size?

How do I use XPath in the search?

Frequently asked questions

Which XPath version does the tool use?

A subset of XPath 1.0: paths, //, attributes, positions, last(), comparisons, contains, starts-with, not, count, and and or. Named axes and union are not included.

How do I query an XML with a namespace?

Write just the local name, such as //numero, and it matches ex:numero. To restrict to one prefix, write the full name, //ex:numero.

Can I export the XPath result?

Click a result to select the node and use the buttons that copy the path and the value. The list shows each hit's full path, ready to paste.