""" :mod:`pandas.io.html` is a module containing functionality for dealing with HTML IO. """ from __future__ import annotations from collections import abc import numbers import re from typing import ( Pattern, Sequence, cast, ) from pandas._typing import ( FilePath, ReadBuffer, ) from pandas.compat._optional import import_optional_dependency from pandas.errors import ( AbstractMethodError, EmptyDataError, ) from pandas.util._decorators import deprecate_nonkeyword_arguments from pandas.core.dtypes.common import is_list_like from pandas.core.construction import create_series_with_explicit_dtype from pandas.core.frame import DataFrame from pandas.io.common import ( file_exists, get_handle, is_url, stringify_path, urlopen, validate_header_arg, ) from pandas.io.formats.printing import pprint_thing from pandas.io.parsers import TextParser _IMPORTS = False _HAS_BS4 = False _HAS_LXML = False _HAS_HTML5LIB = False def _importers() -> None: # import things we need # but make this done on a first use basis global _IMPORTS if _IMPORTS: return global _HAS_BS4, _HAS_LXML, _HAS_HTML5LIB bs4 = import_optional_dependency("bs4", errors="ignore") _HAS_BS4 = bs4 is not None lxml = import_optional_dependency("lxml.etree", errors="ignore") _HAS_LXML = lxml is not None html5lib = import_optional_dependency("html5lib", errors="ignore") _HAS_HTML5LIB = html5lib is not None _IMPORTS = True ############# # READ HTML # ############# _RE_WHITESPACE = re.compile(r"[\r\n]+|\s{2,}") def _remove_whitespace(s: str, regex: Pattern = _RE_WHITESPACE) -> str: """ Replace extra whitespace inside of a string with a single space. Parameters ---------- s : str or unicode The string from which to remove extra whitespace. regex : re.Pattern The regular expression to use to remove extra whitespace. Returns ------- subd : str or unicode `s` with all extra whitespace replaced with a single space. """ return regex.sub(" ", s.strip()) def _get_skiprows(skiprows: int | Sequence[int] | slice | None) -> int | Sequence[int]: """ Get an iterator given an integer, slice or container. Parameters ---------- skiprows : int, slice, container The iterator to use to skip rows; can also be a slice. Raises ------ TypeError * If `skiprows` is not a slice, integer, or Container Returns ------- it : iterable A proper iterator to use to skip rows of a DataFrame. """ if isinstance(skiprows, slice): start, step = skiprows.start or 0, skiprows.step or 1 return list(range(start, skiprows.stop, step)) elif isinstance(skiprows, numbers.Integral) or is_list_like(skiprows): return cast("int | Sequence[int]", skiprows) elif skiprows is None: return 0 raise TypeError(f"{type(skiprows).__name__} is not a valid type for skipping rows") def _read( obj: bytes | FilePath | ReadBuffer[str] | ReadBuffer[bytes], encoding: str | None ) -> str | bytes: """ Try to read from a url, file or string. Parameters ---------- obj : str, unicode, path object, or file-like object Returns ------- raw_text : str """ text: str | bytes if ( is_url(obj) or hasattr(obj, "read") or (isinstance(obj, str) and file_exists(obj)) ): # error: Argument 1 to "get_handle" has incompatible type "Union[str, bytes, # Union[IO[Any], RawIOBase, BufferedIOBase, TextIOBase, TextIOWrapper, mmap]]"; # expected "Union[PathLike[str], Union[str, Union[IO[Any], RawIOBase, # BufferedIOBase, TextIOBase, TextIOWrapper, mmap]]]" with get_handle( obj, "r", encoding=encoding # type: ignore[arg-type] ) as handles: text = handles.handle.read() elif isinstance(obj, (str, bytes)): text = obj else: raise TypeError(f"Cannot read object of type '{type(obj).__name__}'") return text class _HtmlFrameParser: """ Base class for parsers that parse HTML into DataFrames. Parameters ---------- io : str or file-like This can be either a string of raw HTML, a valid URL using the HTTP, FTP, or FILE protocols or a file-like object. match : str or regex The text to match in the document. attrs : dict List of HTML element attributes to match. encoding : str Encoding to be used by parser displayed_only : bool Whether or not items with "display:none" should be ignored Attributes ---------- io : str or file-like raw HTML, URL, or file-like object match : regex The text to match in the raw HTML attrs : dict-like A dictionary of valid table attributes to use to search for table elements. encoding : str Encoding to be used by parser displayed_only : bool Whether or not items with "display:none" should be ignored Notes ----- To subclass this class effectively you must override the following methods: * :func:`_build_doc` * :func:`_attr_getter` * :func:`_text_getter` * :func:`_parse_td` * :func:`_parse_thead_tr` * :func:`_parse_tbody_tr` * :func:`_parse_tfoot_tr` * :func:`_parse_tables` * :func:`_equals_tag` See each method's respective documentation for details on their functionality. """ def __init__( self, io: FilePath | ReadBuffer[str] | ReadBuffer[bytes], match: str | Pattern, attrs: dict[str, str] | None, encoding: str, displayed_only: bool, ): self.io = io self.match = match self.attrs = attrs self.encoding = encoding self.displayed_only = displayed_only def parse_tables(self): """ Parse and return all tables from the DOM. Returns ------- list of parsed (header, body, footer) tuples from tables. """ tables = self._parse_tables(self._build_doc(), self.match, self.attrs) return (self._parse_thead_tbody_tfoot(table) for table in tables) def _attr_getter(self, obj, attr): """ Return the attribute value of an individual DOM node. Parameters ---------- obj : node-like A DOM node. attr : str or unicode The attribute, such as "colspan" Returns ------- str or unicode The attribute value. """ # Both lxml and BeautifulSoup have the same implementation: return obj.get(attr) def _text_getter(self, obj): """ Return the text of an individual DOM node. Parameters ---------- obj : node-like A DOM node. Returns ------- text : str or unicode The text from an individual DOM node. """ raise AbstractMethodError(self) def _parse_td(self, obj): """ Return the td elements from a row element. Parameters ---------- obj : node-like A DOM node. Returns ------- list of node-like These are the elements of each row, i.e., the columns. """ raise AbstractMethodError(self) def _parse_thead_tr(self, table): """ Return the list of thead row elements from the parsed table element. Parameters ---------- table : a table element that contains zero or more thead elements. Returns ------- list of node-like These are the row elements of a table. """ raise AbstractMethodError(self) def _parse_tbody_tr(self, table): """ Return the list of tbody row elements from the parsed table element. HTML5 table bodies consist of either 0 or more elements (which only contain elements) or 0 or more elements. This method checks for both structures. Parameters ---------- table : a table element that contains row elements. Returns ------- list of node-like These are the row elements of a table. """ raise AbstractMethodError(self) def _parse_tfoot_tr(self, table): """ Return the list of tfoot row elements from the parsed table element. Parameters ---------- table : a table element that contains row elements. Returns ------- list of node-like These are the row elements of a table. """ raise AbstractMethodError(self) def _parse_tables(self, doc, match, attrs): """ Return all tables from the parsed DOM. Parameters ---------- doc : the DOM from which to parse the table element. match : str or regular expression The text to search for in the DOM tree. attrs : dict A dictionary of table attributes that can be used to disambiguate multiple tables on a page. Raises ------ ValueError : `match` does not match any text in the document. Returns ------- list of node-like HTML
elements to be parsed into raw data. """ raise AbstractMethodError(self) def _equals_tag(self, obj, tag): """ Return whether an individual DOM node matches a tag Parameters ---------- obj : node-like A DOM node. tag : str Tag name to be checked for equality. Returns ------- boolean Whether `obj`'s tag name is `tag` """ raise AbstractMethodError(self) def _build_doc(self): """ Return a tree-like object that can be used to iterate over the DOM. Returns ------- node-like The DOM from which to parse the table element. """ raise AbstractMethodError(self) def _parse_thead_tbody_tfoot(self, table_html): """ Given a table, return parsed header, body, and foot. Parameters ---------- table_html : node-like Returns ------- tuple of (header, body, footer), each a list of list-of-text rows. Notes ----- Header and body are lists-of-lists. Top level list is a list of rows. Each row is a list of str text. Logic: Use , , elements to identify header, body, and footer, otherwise: - Put all rows into body - Move rows from top of body to header only if all elements inside row are . Move the top all- or while body_rows and row_is_all_th(body_rows[0]): header_rows.append(body_rows.pop(0)) header = self._expand_colspan_rowspan(header_rows) body = self._expand_colspan_rowspan(body_rows) footer = self._expand_colspan_rowspan(footer_rows) return header, body, footer def _expand_colspan_rowspan(self, rows): """ Given a list of s, return a list of text rows. Parameters ---------- rows : list of node-like List of s Returns ------- list of list Each returned row is a list of str text. Notes ----- Any cell with ``rowspan`` or ``colspan`` will have its contents copied to subsequent cells. """ all_texts = [] # list of rows, each a list of str remainder: list[tuple[int, str, int]] = [] # list of (index, text, nrows) for tr in rows: texts = [] # the output for this row next_remainder = [] index = 0 tds = self._parse_td(tr) for td in tds: # Append texts from previous rows with rowspan>1 that come # before this or (see _parse_thead_tr). return row.xpath("./td|./th") def _parse_tables(self, doc, match, kwargs): pattern = match.pattern # 1. check all descendants for the given pattern and only search tables # 2. go up the tree until we find a table xpath_expr = f"//table//*[re:test(text(), {repr(pattern)})]/ancestor::table" # if any table attributes were given build an xpath expression to # search for them if kwargs: xpath_expr += _build_xpath_expr(kwargs) tables = doc.xpath(xpath_expr, namespaces=_re_namespace) tables = self._handle_hidden_tables(tables, "attrib") if self.displayed_only: for table in tables: # lxml utilizes XPATH 1.0 which does not have regex # support. As a result, we find all elements with a style # attribute and iterate them to check for display:none for elem in table.xpath(".//*[@style]"): if "display:none" in elem.attrib.get("style", "").replace(" ", ""): elem.getparent().remove(elem) if not tables: raise ValueError(f"No tables found matching regex {repr(pattern)}") return tables def _equals_tag(self, obj, tag): return obj.tag == tag def _build_doc(self): """ Raises ------ ValueError * If a URL that lxml cannot parse is passed. Exception * Any other ``Exception`` thrown. For example, trying to parse a URL that is syntactically correct on a machine with no internet connection will fail. See Also -------- pandas.io.html._HtmlFrameParser._build_doc """ from lxml.etree import XMLSyntaxError from lxml.html import ( HTMLParser, fromstring, parse, ) parser = HTMLParser(recover=True, encoding=self.encoding) try: if is_url(self.io): with urlopen(self.io) as f: r = parse(f, parser=parser) else: # try to parse the input in the simplest way r = parse(self.io, parser=parser) try: r = r.getroot() except AttributeError: pass except (UnicodeDecodeError, OSError) as e: # if the input is a blob of html goop if not is_url(self.io): r = fromstring(self.io, parser=parser) try: r = r.getroot() except AttributeError: pass else: raise e else: if not hasattr(r, "text_content"): raise XMLSyntaxError("no text parsed from document", 0, 0, 0) return r def _parse_thead_tr(self, table): rows = [] for thead in table.xpath(".//thead"): rows.extend(thead.xpath("./tr")) # HACK: lxml does not clean up the clearly-erroneous # . (Missing ). Add # the and _pretend_ it's a ; _parse_td() will find its # children as though it's a . # # Better solution would be to use html5lib. elements_at_root = thead.xpath("./td|./th") if elements_at_root: rows.append(thead) return rows def _parse_tbody_tr(self, table): from_tbody = table.xpath(".//tbody//tr") from_root = table.xpath("./tr") # HTML spec: at most one of these lists has content return from_tbody + from_root def _parse_tfoot_tr(self, table): return table.xpath(".//tfoot//tr") def _expand_elements(body): data = [len(elem) for elem in body] lens = create_series_with_explicit_dtype(data, dtype_if_empty=object) lens_max = lens.max() not_max = lens[lens != lens_max] empty = [""] for ind, length in not_max.items(): body[ind] += empty * (lens_max - length) def _data_to_frame(**kwargs): head, body, foot = kwargs.pop("data") header = kwargs.pop("header") kwargs["skiprows"] = _get_skiprows(kwargs["skiprows"]) if head: body = head + body # Infer header when there is a or top
- Move rows from bottom of body to footer only if all elements inside row are """ header_rows = self._parse_thead_tr(table_html) body_rows = self._parse_tbody_tr(table_html) footer_rows = self._parse_tfoot_tr(table_html) def row_is_all_th(row): return all(self._equals_tag(t, "th") for t in self._parse_td(row)) if not header_rows: # The table has no
rows from # body_rows to header_rows. (This is a common case because many # tables in the wild have no
while remainder and remainder[0][0] <= index: prev_i, prev_text, prev_rowspan = remainder.pop(0) texts.append(prev_text) if prev_rowspan > 1: next_remainder.append((prev_i, prev_text, prev_rowspan - 1)) index += 1 # Append the text from this , colspan times text = _remove_whitespace(self._text_getter(td)) rowspan = int(self._attr_getter(td, "rowspan") or 1) colspan = int(self._attr_getter(td, "colspan") or 1) for _ in range(colspan): texts.append(text) if rowspan > 1: next_remainder.append((index, text, rowspan - 1)) index += 1 # Append texts from previous rows at the final position for prev_i, prev_text, prev_rowspan in remainder: texts.append(prev_text) if prev_rowspan > 1: next_remainder.append((prev_i, prev_text, prev_rowspan - 1)) all_texts.append(texts) remainder = next_remainder # Append rows that only appear because the previous row had non-1 # rowspan while remainder: next_remainder = [] texts = [] for prev_i, prev_text, prev_rowspan in remainder: texts.append(prev_text) if prev_rowspan > 1: next_remainder.append((prev_i, prev_text, prev_rowspan - 1)) all_texts.append(texts) remainder = next_remainder return all_texts def _handle_hidden_tables(self, tbl_list, attr_name): """ Return list of tables, potentially removing hidden elements Parameters ---------- tbl_list : list of node-like Type of list elements will vary depending upon parser used attr_name : str Name of the accessor for retrieving HTML attributes Returns ------- list of node-like Return type matches `tbl_list` """ if not self.displayed_only: return tbl_list return [ x for x in tbl_list if "display:none" not in getattr(x, attr_name).get("style", "").replace(" ", "") ] class _BeautifulSoupHtml5LibFrameParser(_HtmlFrameParser): """ HTML to DataFrame parser that uses BeautifulSoup under the hood. See Also -------- pandas.io.html._HtmlFrameParser pandas.io.html._LxmlFrameParser Notes ----- Documentation strings for this class are in the base class :class:`pandas.io.html._HtmlFrameParser`. """ def __init__(self, *args, **kwargs): super().__init__(*args, **kwargs) from bs4 import SoupStrainer self._strainer = SoupStrainer("table") def _parse_tables(self, doc, match, attrs): element_name = self._strainer.name tables = doc.find_all(element_name, attrs=attrs) if not tables: raise ValueError("No tables found") result = [] unique_tables = set() tables = self._handle_hidden_tables(tables, "attrs") for table in tables: if self.displayed_only: for elem in table.find_all(style=re.compile(r"display:\s*none")): elem.decompose() if table not in unique_tables and table.find(string=match) is not None: result.append(table) unique_tables.add(table) if not result: raise ValueError(f"No tables found matching pattern {repr(match.pattern)}") return result def _text_getter(self, obj): return obj.text def _equals_tag(self, obj, tag): return obj.name == tag def _parse_td(self, row): return row.find_all(("td", "th"), recursive=False) def _parse_thead_tr(self, table): return table.select("thead tr") def _parse_tbody_tr(self, table): from_tbody = table.select("tbody tr") from_root = table.find_all("tr", recursive=False) # HTML spec: at most one of these lists has content return from_tbody + from_root def _parse_tfoot_tr(self, table): return table.select("tfoot tr") def _setup_build_doc(self): raw_text = _read(self.io, self.encoding) if not raw_text: raise ValueError(f"No text parsed from document: {self.io}") return raw_text def _build_doc(self): from bs4 import BeautifulSoup bdoc = self._setup_build_doc() if isinstance(bdoc, bytes) and self.encoding is not None: udoc = bdoc.decode(self.encoding) from_encoding = None else: udoc = bdoc from_encoding = self.encoding return BeautifulSoup(udoc, features="html5lib", from_encoding=from_encoding) def _build_xpath_expr(attrs) -> str: """ Build an xpath expression to simulate bs4's ability to pass in kwargs to search for attributes when using the lxml parser. Parameters ---------- attrs : dict A dict of HTML attributes. These are NOT checked for validity. Returns ------- expr : unicode An XPath expression that checks for the given HTML attributes. """ # give class attribute as class_ because class is a python keyword if "class_" in attrs: attrs["class"] = attrs.pop("class_") s = " and ".join([f"@{k}={repr(v)}" for k, v in attrs.items()]) return f"[{s}]" _re_namespace = {"re": "http://exslt.org/regular-expressions"} class _LxmlFrameParser(_HtmlFrameParser): """ HTML to DataFrame parser that uses lxml under the hood. Warning ------- This parser can only handle HTTP, FTP, and FILE urls. See Also -------- _HtmlFrameParser _BeautifulSoupLxmlFrameParser Notes ----- Documentation strings for this class are in the base class :class:`_HtmlFrameParser`. """ def _text_getter(self, obj): return obj.text_content() def _parse_td(self, row): # Look for direct children only: the "row" element here may be a #
foobar
-only rows if header is None: if len(head) == 1: header = 0 else: # ignore all-empty-text rows header = [i for i, row in enumerate(head) if any(text for text in row)] if foot: body += foot # fill out elements of body that are "ragged" _expand_elements(body) with TextParser(body, header=header, **kwargs) as tp: return tp.read() _valid_parsers = { "lxml": _LxmlFrameParser, None: _LxmlFrameParser, "html5lib": _BeautifulSoupHtml5LibFrameParser, "bs4": _BeautifulSoupHtml5LibFrameParser, } def _parser_dispatch(flavor: str | None) -> type[_HtmlFrameParser]: """ Choose the parser based on the input flavor. Parameters ---------- flavor : str The type of parser to use. This must be a valid backend. Returns ------- cls : _HtmlFrameParser subclass The parser class based on the requested input flavor. Raises ------ ValueError * If `flavor` is not a valid backend. ImportError * If you do not have the requested `flavor` """ valid_parsers = list(_valid_parsers.keys()) if flavor not in valid_parsers: raise ValueError( f"{repr(flavor)} is not a valid flavor, valid flavors are {valid_parsers}" ) if flavor in ("bs4", "html5lib"): if not _HAS_HTML5LIB: raise ImportError("html5lib not found, please install it") if not _HAS_BS4: raise ImportError("BeautifulSoup4 (bs4) not found, please install it") # Although we call this above, we want to raise here right before use. bs4 = import_optional_dependency("bs4") # noqa:F841 else: if not _HAS_LXML: raise ImportError("lxml not found, please install it") return _valid_parsers[flavor] def _print_as_set(s) -> str: arg = ", ".join([pprint_thing(el) for el in s]) return f"{{{arg}}}" def _validate_flavor(flavor): if flavor is None: flavor = "lxml", "bs4" elif isinstance(flavor, str): flavor = (flavor,) elif isinstance(flavor, abc.Iterable): if not all(isinstance(flav, str) for flav in flavor): raise TypeError( f"Object of type {repr(type(flavor).__name__)} " f"is not an iterable of strings" ) else: msg = repr(flavor) if isinstance(flavor, str) else str(flavor) msg += " is not a valid flavor" raise ValueError(msg) flavor = tuple(flavor) valid_flavors = set(_valid_parsers) flavor_set = set(flavor) if not flavor_set & valid_flavors: raise ValueError( f"{_print_as_set(flavor_set)} is not a valid set of flavors, valid " f"flavors are {_print_as_set(valid_flavors)}" ) return flavor def _parse(flavor, io, match, attrs, encoding, displayed_only, **kwargs): flavor = _validate_flavor(flavor) compiled_match = re.compile(match) # you can pass a compiled regex here retained = None for flav in flavor: parser = _parser_dispatch(flav) p = parser(io, compiled_match, attrs, encoding, displayed_only) try: tables = p.parse_tables() except ValueError as caught: # if `io` is an io-like object, check if it's seekable # and try to rewind it before trying the next parser if hasattr(io, "seekable") and io.seekable(): io.seek(0) elif hasattr(io, "seekable") and not io.seekable(): # if we couldn't rewind it, let the user know raise ValueError( f"The flavor {flav} failed to parse your input. " "Since you passed a non-rewindable file " "object, we can't rewind it to try " "another parser. Try read_html() with a different flavor." ) from caught retained = caught else: break else: assert retained is not None # for mypy raise retained ret = [] for table in tables: try: ret.append(_data_to_frame(data=table, **kwargs)) except EmptyDataError: # empty table continue return ret @deprecate_nonkeyword_arguments(version="2.0") def read_html( io: FilePath | ReadBuffer[str], match: str | Pattern = ".+", flavor: str | None = None, header: int | Sequence[int] | None = None, index_col: int | Sequence[int] | None = None, skiprows: int | Sequence[int] | slice | None = None, attrs: dict[str, str] | None = None, parse_dates: bool = False, thousands: str | None = ",", encoding: str | None = None, decimal: str = ".", converters: dict | None = None, na_values=None, keep_default_na: bool = True, displayed_only: bool = True, ) -> list[DataFrame]: r""" Read HTML tables into a ``list`` of ``DataFrame`` objects. Parameters ---------- io : str, path object, or file-like object String, path object (implementing ``os.PathLike[str]``), or file-like object implementing a string ``read()`` function. The string can represent a URL or the HTML itself. Note that lxml only accepts the http, ftp and file url protocols. If you have a URL that starts with ``'https'`` you might try removing the ``'s'``. match : str or compiled regular expression, optional The set of tables containing text matching this regex or string will be returned. Unless the HTML is extremely simple you will probably need to pass a non-empty string here. Defaults to '.+' (match any non-empty string). The default value will return all tables contained on a page. This value is converted to a regular expression so that there is consistent behavior between Beautiful Soup and lxml. flavor : str, optional The parsing engine to use. 'bs4' and 'html5lib' are synonymous with each other, they are both there for backwards compatibility. The default of ``None`` tries to use ``lxml`` to parse and if that fails it falls back on ``bs4`` + ``html5lib``. header : int or list-like, optional The row (or list of rows for a :class:`~pandas.MultiIndex`) to use to make the columns headers. index_col : int or list-like, optional The column (or list of columns) to use to create the index. skiprows : int, list-like or slice, optional Number of rows to skip after parsing the column integer. 0-based. If a sequence of integers or a slice is given, will skip the rows indexed by that sequence. Note that a single element sequence means 'skip the nth row' whereas an integer means 'skip n rows'. attrs : dict, optional This is a dictionary of attributes that you can pass to use to identify the table in the HTML. These are not checked for validity before being passed to lxml or Beautiful Soup. However, these attributes must be valid HTML table attributes to work correctly. For example, :: attrs = {'id': 'table'} is a valid attribute dictionary because the 'id' HTML tag attribute is a valid HTML attribute for *any* HTML tag as per `this document `__. :: attrs = {'asdf': 'table'} is *not* a valid attribute dictionary because 'asdf' is not a valid HTML attribute even if it is a valid XML attribute. Valid HTML 4.01 table attributes can be found `here `__. A working draft of the HTML 5 spec can be found `here `__. It contains the latest information on table attributes for the modern web. parse_dates : bool, optional See :func:`~read_csv` for more details. thousands : str, optional Separator to use to parse thousands. Defaults to ``','``. encoding : str, optional The encoding used to decode the web page. Defaults to ``None``.``None`` preserves the previous encoding behavior, which depends on the underlying parser library (e.g., the parser library will try to use the encoding provided by the document). decimal : str, default '.' Character to recognize as decimal point (e.g. use ',' for European data). converters : dict, default None Dict of functions for converting values in certain columns. Keys can either be integers or column labels, values are functions that take one input argument, the cell (not column) content, and return the transformed content. na_values : iterable, default None Custom NA values. keep_default_na : bool, default True If na_values are specified and keep_default_na is False the default NaN values are overridden, otherwise they're appended to. displayed_only : bool, default True Whether elements with "display: none" should be parsed. Returns ------- dfs A list of DataFrames. See Also -------- read_csv : Read a comma-separated values (csv) file into DataFrame. Notes ----- Before using this function you should read the :ref:`gotchas about the HTML parsing libraries `. Expect to do some cleanup after you call this function. For example, you might need to manually assign column names if the column names are converted to NaN when you pass the `header=0` argument. We try to assume as little as possible about the structure of the table and push the idiosyncrasies of the HTML contained in the table to the user. This function searches for ```` elements and only for ```` and ```` or ```` argument, it is used to construct the header, otherwise the function attempts to find the header within the body (by putting rows with only ``
`` rows and ```` elements within each ``
`` element in the table. ```` stands for "table data". This function attempts to properly handle ``colspan`` and ``rowspan`` attributes. If the function has a ``
`` elements into the header). Similar to :func:`~read_csv` the `header` argument is applied **after** `skiprows` is applied. This function will *always* return a list of :class:`DataFrame` *or* it will fail, e.g., it will *not* return an empty list. Examples -------- See the :ref:`read_html documentation in the IO section of the docs ` for some examples of reading in HTML tables. """ _importers() # Type check here. We don't want to parse only to fail because of an # invalid value of an integer skiprows. if isinstance(skiprows, numbers.Integral) and skiprows < 0: raise ValueError( "cannot skip rows starting from the end of the " "data (you passed a negative value)" ) validate_header_arg(header) io = stringify_path(io) return _parse( flavor=flavor, io=io, match=match, header=header, index_col=index_col, skiprows=skiprows, parse_dates=parse_dates, thousands=thousands, attrs=attrs, encoding=encoding, decimal=decimal, converters=converters, na_values=na_values, keep_default_na=keep_default_na, displayed_only=displayed_only, )