用等价单词代替句子

Question

I would like to change sentences with equivalent words; For example:

Reference Database:

put <-> set
returns <-> switch
between <-> among
paragraphs <-> null
between paragraphs <-> among paragraphs

Sentence:

put returns between paragraphs

Replaced Sentence:

set switch among paragraphs

Yes; finding and replacing easy and fine. But the question: I have " paragraphs " and " between paragraphs " in records. How can I search and replace long text before shorts? Performance is very important because might reference database has over 1 million rows.

Currently using: Entity Framework, C#, MVC, SQL Server 2014

Any help will be fantastic. Thanks.

Update:

Sentence:

string str = "The number of cases in the Ebola outbreak passes 10,000, with 4,922 deaths, the World Health Organization's latest report says."

And Database has 1,000,000 records; I can do it with get all rows from database, and foreach all to text. This is stupid method. I need a way to sentences to database How can I select rows with sentences?

I have a method like split all text with ( ) white-space and then try to select in database. But it'll pass long texts with white-space. "The number of cases"; it will search it like "the", "number", "of", "cases" with seperated search... And records will not useful.

Answer 1

You could organise your replacement database in a trie-like structure . All single-word expressions are on the first level. Multi-word expressions are stored as descendants of preceding words in the expression. In your example:

root
    -> put: set
    -> returns: switch
    -> between: among
        -> paragraphs: among paragraphs
    -> paragraphs: sections

The root will be a dictionary of words. Each node has a sub-dictionary, which will be null in most cases, and a replacement value. The replacement may be null in intermediate words, for example in ìn -> this -> case , this won't have a replacement, because in -> this isn't a valid replacement in itself.

Split your sentence and iterate though the words. If you find a possible start of a replacement, follow the trie and determine the longest possible replacement at this position. Replace that and continue the iteration from the next word.

 between you and me -> among you and me
 between other paragraphs -> among other sections
 between paragraphs -> among paragraphs

When you split the words, keep the space and punctuation between the words and follow the trie nodes only if there is no punctuation between the words, so that a sentence like

 There must be something in between; paragraphs 1 and 2 seem to indicate that.

is treated correctly.

This method should be efficient if you have many sentences that need to be replaced with replacements from the same database. The database trie must be built only once. If you have only a few sentences to replace or if your database changes frequently, this is not a good approach.

Answer 2

you can use REPLACE function in sql server

SELECT REPLACE('abcdefghicde','cde','xxx') GO

this replace cde with xxx

in your application the sql maybe

UPDATE tablename set col=REPLACE(col,' put ',' set ') //Please notice there are 2 ' ' beside the word