Skip to content
zhejiangxiaomai edited this page Dec 6, 2015 · 37 revisions

CLAIMS DML : LOAD,INSERT,SELECT

1. LOAD

(This statement is used to import data from the disk)

  1. Format

    LOAD TABLE table_name FROM expr_list with STRING, STRING;
    
  2. Explanation

    expr_list:Specify the import data file path, you can enter multiple file paths, separated by commas. eg.“/home/imdb/data/table1”,”/home/imdb/data/table2”.

    The first column delimiter STRING specify only a character, if more than one character, take the first character. eg. ’\t’。 The second column delimiter STRING specify only a character, if more than one character, take the first character. eg. ’\n’。

  3. Example

LOAD TABLE a from “/home/imdb/table.csv” with ‘\t’,’\n’;

2. INSERT

(This statement is used to insert one or more data within a particular table)

  1. Format

    INSERT [INTO] table_name [(column_name, …)] VALUES (expr, …), …;
    
  2. Explanation

    column_name:Specify the insertion of data for the column, if not, the default insert all columns.
    (Expr,…):Specify the insertion of content, you need to correspond with the specified column.Insert              multiple columns of data can use (expr, …), …,Each data included by the parentheses, separated by commas
    
  3. Example

    insert into test1 values(10,1.11);

    insert into region values(4,'Middle'),(3,'Europe'),(2,'Asia'),(1,'America'),(0,'Africa');

3. SELECT

  1. Format

      SELECT [ALL | DISTINCT] select_list [AS other_name]  
      FROM table_name | join_table
      [WHERE where_conditions]
      [GROUP BY group_by_list]
      [HAVING search_conditions] 
      [ORDER BY order_list [ASC | DESC]]
      [LIMIT [offset,]row_count ];
    
  2. SELECT Clause Description

Clause Explanation notation
ALL | DISTINCT In the database table may contain duplicate values. Designated "DISTINCT", in the same row on the query results show only one line; specify "ALL", then list all of the line; not specified, the default is "ALL". Does not support now
select_list Lists the column name or expression to query, use the "," separated,use "*" present all columns See more details about expression in 3.3 Expression Support Sheet
AS other_name Optionally, rename the output field Aliases can be used in the having and order by
FROM table_name | join_table Required,specify the table which be read data,support connection between tables bo not support 'using' in join
WHERE where_conditions optional, WHERE clause for set a shuffle condition query results only contain the tuple meet the conditions
GROUP BY group_by_list for Subtotals If it is expression, and appear in the select list, then the expression must have aliases, and must refer to the alias in orderby, having
HAVING search_conditions HAVING clause is similar to the WHERE clause, but the HAVING clause can use aggregate functions (such as SUM, AVG, etc.), the prerequisite is to have a GROUP By clause appears.
ORDER BY order_list [ASC | DESC] For ascending (ASC) or descending (DESC) display the query results. When you do not specify ASC or DESC, the default is ASC. order_list only specific properties, does not support digital
[LIMIT [offset,] row_count ] Forced SELECT statement returns the specified number of records. LIMIT accept one or two digits natural numbers constant parameters, the offset (the default is 0), the maximum number

3.Expression Support Sheet

Expression Type Parameters required Example Remark
expr + expr , expr - expr , expr * expr , expr / expr , expr % expr Compatible 1+2 , 2-1 , 3*4 , 4/2 , 4%2 do not boolean type data
expr AND expr , expr OR expr Numeric, logic 1 and 0 , 1 or 0
expr CMP (=,!=,>,<,>=,<=) expr Compatible a>b
Expr[NOT] IN const_list Corresponding parameter type is compatible a in (2,4,5) IN=(=ANY)
expr_row [NOT] IN const_row_list Corresponding parameter type is compatible (a,b) in ((1,'a'),(2,'b'))
- expr(negative expression) the type has negative -a
NOT expr(Negative logic), ! expr(Negative logic)=NOT expr logic not a , ! a
SUBSTRING(expr,start_position,length) start_position,length is Non-negative integer substring('string',2,4)
TRIM([{BOTH | LEADING | TRAILING} [remstr] FROM] str) remstr is a string which length is 1 trim(' abc '); , trim(BOTH 'a' FROM 'abca'); LEADING Start from the head , TRAILING Start from the tail
UPPER(expr) String upper('abc')
expr [NOT]LIKE expr1 String 'abc' like '_b%' "_" present one char,"%" present any chars
CASE WHEN expr1 then expr2 [WHEN expr1 then expr2 else expr3] END The return type of expr1 is logical, expr2 is String case when a%2=0 then 'even' when a%2=1 then 'odd' else 'none' end
CASE expr1 WHEN expr2 THEN expr3 [WHEN expr2 THEN expr3 else expr4] END "=" is compatible in expr1 and expr2 ,expr3 and expr4 is string case sex when 1 then ‘male’ else ‘female’ end it means expr1=expr2,so present expr3
expr BETWEEN expr1 AND expr2 expr1 and expr2 's compare operation is compatible with the expr a between 2 and 4
DATE_ADD(date,INTERVAL expr type) DATE_SUB(date,INTERVAL expr type) date : Date type String , expr : non-negative Integer , type : day,month,year DATE_ADD('2014-9-7',INTERVAL 3 day) Date constants must be delimited as 2010-10-05

4.Special Case Description

  •      Region Table
    
regionkey (bigint unsigned) name(varchar(30))
0 Africa
1 America
2 Asia
3 Europe
4 Middle
  •      National table
    
nationkey (bigint unsigned) name(varchar(30)) regionkey (bigint unsigned)
0 CHINA 2
1 UNITED STATES 1
2 RUSSIA 3
3 KENYA 0
4 JORDAN 4
5 INDIA 2

(1) aggregation function and group by

if "select_list" exist aggregation function(SUM(),COUNT(),AVG(),MAX(),MIN() and COUNT(*), Then the non-aggregation function attribute columns exist in "select_list" are a subset of the attribute columns in "groupby_list".If a expression in "groupby_list",and it also appear in "select_list",then this expression must use alias in "select_list",and use this alias in "orderby" and "having" refer to this expression. "groupby" expression mathematical operation.

select max(nationkey),regionkey%5 as regionkey from nation group by regionkey;

Result

max(nationkey) regionkey
3 0
1 1
5 2
2 3
4 4
Explanation: regionkey%5 must declare a alias in select_list like regionkey, in group by use regionkey to refer.

(2)join

join connection can be divided into inner join and outer join.outer join can be divided into left outer join and right outer join and full outer join.After the two tables join, you can use "ON specified condition"to filter.Currently ,CLAIMS support " Ta join Tb [ON condition]" support equal join and corss join。

select nation.name,region.name from nation join region on nation.regionkey = region.regionkey;

Result

nation.name region.name
CHINA Asia
UNITED STATES America
RUSSIA Europe
KENYA Africa
JORDAN Middle
INDIA Asia

Clone this wiki locally