If you want to learn how to create and manage a Translation Memory, please refer to this article.
If you want to learn how to add new segments to the Translation Memories, please refer to this one.
This article will show you how to use the Translation Memory search and find terms or segments in an assertive way.
Once you find the Translation Memory you want to look into, you can start your search. Our platform provides the following search options on the TMs:
Full Text Search;
Fuzzy Search using Threshold;
Contains Search, which can use RegEx.
You can adjust the number of Rows, which refer to how many rows of results you want to visualize on the same page. As for the Columns, they refer to which Target Languages you want to show on your search.
You can also select from which language you want to search, either from the source language or from one of the targets.
Full Text Search
The Full Text Search will return any segment that contains the text that is being prompted.
Additionally, the Full Text Search will tokenize the word that is being prompted, and a Stemming Process will be applied. Stemming is the process of reducing a word to its root form, or stem.The idea is to treat different word forms with the same base meaning as equivalent during a search. For instance, if we look for the word "testing", it will bring the related results:
Fuzzy Search
The Fuzzy Search will take into consideration the entire segment. The lowest Threshold you can select is 49%.
Given the following segment "we are stressing this test" that is already saved in the Translation Memory. If you type exactly the same segment, your result will show a 100% match:
If you type only the following and keep the threshold as 49%, the search will show you every find with a 49% match or upper. In this case, we have an 61.54% match.
If the threshold was defined as 75%, the segment would not appear.
The Fuzzy search will always respond according to the percentage of the segment you type and the Threshold you select.
Contains Search
The Contains Search will work like the Full Text Search, but it will not tokenize nor will it apply any stemming process. The Contains Search will look up for the exact word that is being searched. Additionally, you can enable the RegEx search:
With the RegEx search enabled, you can perform more specific searches. For example, if you look for .*test.*, it will bring every segment that contains the test word.
However, if we insert whitespaces between the operators, the search will only retrieve the segments that have the test word between whitespaces, excluding the ones that are between quotes.
It goes the same way around. If we look for 'test', between the apostrophes, it will only retrieve the entries that match exactly the term that has been searched for, not including the other ones that are not between the defined punctuation signs.
