The (character) string or simply the character data type, how is named by the MSDN documentation, is one of the primary data types available in a database and probably the most complex given the fact that it can encompass any combination of numeric, literal, date or any other primary data type. In fact it could include any type of chunk of text that could be written in any language as SQL Server supports Unicode and thus most of the (written) languages. In this post I’m not intending to make a complete anthology of strings, and neither to retake the whole volume of information available in MSDN or other important books. My intent is to look at strings from a slightly different perspective, considering the various functions involving strings and how they could be used in order to provide various functionality, in fact the cornerstone of everyday developer.
A few things to remember:
1. there were mainly two types of non-unicode strings: the char (or character) of fixed length, respectively the varchar of variable length (varying character)
2. if initially both types of strings were having a maximum length of 8000 of characters, with SQL Server it’s possible to have a varchar with maximum storage size, declared as varchar(max).
3. if in the past there were some strict recommendations in what concerns the use of char or varchar, nowadays the varchar tends to be used almost everywhere, even for the strings of length 1.
4. talking about length, it denotes the number of chracters a string stores.
5. the trailing spaces, the spaces found at the right extremity of a string are typically ignored, while the leading spaces, the spaces found at the left extremity, are not ignored.
6. starting with SQL Server 2000, for the two character data types were introduced the corresponding unicode data types prefixed with n (national): nchar, respectively nvarchar.
7. given the fact that a unicode character needs more space to store the same non-unicode string, actually the number of bits doubles, the maximum length for an unicode string is only 4000.
8. there is also a non-unicode text, respectively ntext unicode data type, designed to store maximum length, though as it seems they could become soon deprecated, so might be a good idea to avoid it.
9. not-initialized variables, including strings, have the value NULL, referred also the NULL string, therefore it’s always a good idea to initialize your variables.
10. by empty string string is designated the string containing no character “’’”, and has the length 0.
11. there are several specific functions available for the creation, manipulation and conversion of strings from and to other data types.
12. not all of the aggregated functions work with string values (specifically the ones requesting a number value like SUM, AVG, STDV).
13. the operations performed on strings of different data types are generally not impacted by this aspect, though there are some exceptions.
14. there are several (basic) operations with strings, typically concatenation, extraction of subpart of a string, insertion, replacement or deletion of characters, rearangement of string’s characters, trimming, splitting it in substrings (decomposition), etc.
15. there are several numeric values based on strings: length, position of a given text in a text, number of not empty characters, encoding of a character, on whether the text represents a valid numeric or date values, etc.
Please note that this is work in progress so take it as such!!! I’m trying to put together things learned long time ago, so maybe I will have to reformulate and add new things on the list. Any additions are more than welcome!