Is there a less bug-prone way to include quote in selector string variable when defining opt.TableSelector?

Question

0 个投票

My situation is to readtable( ) several tables, each of which has a specific title, from an html file. I have come up with a couple solutions to define TableSelector, but it is bug prone. Is there a better way?

opt = htmlImportOptions;
opt.TableSelector = "//TABLE[contains(.,'Cash dividends')]";
headings = ["'Music'", "Photo"];
%% correct selector string, but bug prone
opt.TableSelector = "//TABLE[contains(.," + headings(1) +")]";
opt.TableSelector
ans = "//TABLE[contains(.,'Music')]"
% readtable(html, opt) 
%% wrong solution
opt.TableSelector = "TABLE[contains(.,"+ headings(2) +")]";
opt.TableSelector
ans = "TABLE[contains(.,Photo)]"
% readtable(html, opt) % would return no error message, so it's a bug source.
%% fix the wrong solution
str1 = "TABLE[contains(.,'";
str2 = "')]";
opt.TableSelector = str1 + headings(2) + str2;
opt.TableSelector
ans = "TABLE[contains(.,'Photo')]"
% readtable(html, opt)

3 个评论
显示 1更早的评论隐藏 1更早的评论

Walter Roberson 2023-5-15

Reading further, I see that we do not need to worry about "escape characters" -- for example if the user specified 'Music\' then the \' would be treated as two distinct unrelated characters, not as indicating "the next character is a literal quote". So we do not need to "sanitize" the input against an assortment of characters.

We do, however, potentially need to worry about the possibility that the headings contain ' characters or " characters.

https://stackoverflow.com/questions/14822153/escape-single-quote-in-xpath-with-nokogiri