Splunk SPL for SQL users
This is not a perfect mapping between SQL and Splunk Search Processing Language (SPL), but if you are familiar with SQL, this quick comparison might be helpful as a jump-start into using the search commands.
Concepts
The Splunk platform does not store data in a conventional database. Rather, it stores data in a distributed, non-relational, semi-structured database with an implicit time dimension. Relational databases require that all table columns be defined up-front and they do not automatically scale by just plugging in new hardware. However, there are analogues to many of the concepts in the database world.
Database Concept | Splunk Concept | Notes |
---|---|---|
SQL query | Splunk search | A Splunk search retrieves indexed data and can perform transforming and reporting operations. Results from one search can be "piped", or transferred, from command to command, to filter, modify, reorder, and group your results. |
table/view | search results | Search results can be thought of as a database view, a dynamically generated table of rows, with columns. |
index | index | All values and fields are indexed by Splunk software, so there is no need to manually add, update, drop, or even think about indexing columns. Everything can be quickly retrieved automatically. |
row | result/event | A result in a Splunk search is a list of fields (i.e., column) values, corresponding to a table row. An event is a result that has a timestamp and raw text. Typically an event is a record from a log file, such as:
|
column | field | Fields are returned dynamically from a search, meaning that one search might return a set of fields, while another search might return another set. After teaching Splunk software how to extract more fields from the raw underlying data, the same search will return more fields than it previously did. Fields are not tied to a datatype. |
database/schema | index/app | A Splunk index is a collection of data, somewhat like a database has a collection of tables. Domain knowledge of that data, how to extract it, what reports to run, etc, are stored in a Splunk application. |
From SQL to Splunk SPL
SQL is designed to search relational database tables which are comprised of columns. SPL is designed to search events, which are comprised of fields. In SQL, you often see examples that use "mytable" and "mycolumn". In SPL, you will see examples that refer to "fields". In these examples, the "source" field is used as a proxy for "table". In Splunk software, "source" is the name of the file, stream, or other input from which a particular piece of data originates, for example /var/log/messages
or UDP:514
.
When translating from any language to another, often the translation is longer because of idioms in the original language. Some of the Splunk search examples shown below could be more concise and more efficient, but for parallelism and clarity, the SPL table and field names are kept the same as the SQL example.
- SPL searches rarely need the FIELDS command to filter out columns because the user interface provides a more convenient method for filtering. The FIELDS command is used in the SPL examples for parallelism.
- With SPL, you never have to use the AND operator in Boolean searches, because AND is implied between terms. However when you use the AND or OR operators, they must be specified in uppercase.
- SPL commands do not need to be specified in uppercase. In the these SPL examples, the commands are specified in uppercase for easier identification and clarity.
- Although some SPL commands loosely correspond to specific SQL commands as shown in the following table, your SPL searches might not produce the desired results if you "think in SQL." For this reason, avoid directly translating from SQL to SPL when you design your searches. See About the search language in the Search Manual for an overview of SPL.
SQL command | SQL example | Splunk SPL example |
---|---|---|
SELECT * |
SELECT * FROM mytable |
source=mytable |
WHERE | SELECT * FROM mytable WHERE mycolumn=5 |
source=mytable mycolumn=5 |
SELECT | SELECT mycolumn1, mycolumn2 FROM mytable |
source=mytable | FIELDS mycolumn1, mycolumn2 |
AND/OR | SELECT * FROM mytable WHERE (mycolumn1="true" OR mycolumn2="red") AND mycolumn3="blue" |
source=mytable AND (mycolumn1="true" OR mycolumn2="red") AND mycolumn3="blue" Note: The AND operator is implied in SPL and does not need to be specified. For this example you could also use: source=mytable (mycolumn1="true" OR mycolumn2="red") mycolumn3="blue" |
AS (alias) | SELECT mycolumn AS column_alias FROM mytable |
source=mytable | RENAME mycolumn as column_alias | FIELDS column_alias |
BETWEEN | SELECT * FROM mytable WHERE mycolumn BETWEEN 1 AND 5 |
source=mytable mycolumn>=1 mycolumn<=5 |
GROUP BY | SELECT mycolumn, avg(mycolumn) FROM mytable WHERE mycolumn=value GROUP BY mycolumn |
source=mytable mycolumn=value | STATS avg(mycolumn) BY mycolumn | FIELDS mycolumn, avg(mycolumn) Several commands use a |
HAVING | SELECT mycolumn, avg(mycolumn) FROM mytable WHERE mycolumn=value GROUP BY mycolumn HAVING avg(mycolumn)=value |
source=mytable mycolumn=value | STATS avg(mycolumn) BY mycolumn | SEARCH avg(mycolumn)=value | FIELDS mycolumn, avg(mycolumn) |
LIKE | SELECT * FROM mytable WHERE mycolumn LIKE "%some text%" |
source=mytable mycolumn="*some text*" Note: The most common search in Splunk SPL is nearly impossible in SQL - to search all fields for a substring. The following SPL search returns all rows that contain "some text" anywhere: source=mytable "some text" |
ORDER BY | SELECT * FROM mytable ORDER BY mycolumn desc |
source=mytable | SORT -mycolumn In SPL you use a negative sign ( - ) in front of a field name to sort in descending order. |
SELECT DISTINCT | SELECT DISTINCT mycolumn1, mycolumn2 FROM mytable |
source=mytable | DEDUP mycolumn1, mycolumn2 | FIELDS mycolumn1, mycolumn2 |
SELECT TOP | SELECT TOP(5) mycolum1, mycolum2 FROM mytable1 WHERE mycolum3 = "bar" ORDER BY mycolum1 mycolum2 |
Source=mytable1 mycolum3="bar" | FIELDS mycolum1 mycolum2 | SORT mycolum1 mycolum2 | HEAD 5 |
INNER JOIN | SELECT * FROM mytable1 INNER JOIN mytable2 ON mytable1.mycolumn= mytable2.mycolumn |
index=myIndex1 OR index=myIndex2 | stats values(*) AS * BY myField Note: There are two other methods to join tables:
... | LOOKUP myvaluelookup mycolumn OUTPUT myoutputcolumn
source=mytable1 [SEARCH source=mytable2 mycolumn2=myvalue | FIELDS mycolumn2] If the columns that you want to join on have different names, use the source=mytable1 | JOIN type=inner mycolumn [ SEARCH source=mytable2 | RENAME mycolumn2 AS mycolumn] To rename the column in myindex1: index=myIndex1 OR index=myIndex2 | rename myfield1 as myField | stats values(*) AS * BY myField You can rename a column regardless of whether you use the search command, a lookup, or a subsearch. |
LEFT (OUTER) JOIN | SELECT * FROM mytable1 LEFT JOIN mytable2 ON mytable1.mycolumn= mytable2.mycolumn |
source=mytable1 | JOIN type=left mycolumn [SEARCH source=mytable2] |
SELECT INTO | SELECT * INTO new_mytable IN mydb2 FROM old_mytable |
source=old_mytable | EVAL source=new_mytable | COLLECT index=mydb2 Note: COLLECT is typically used to store expensively calculated fields back into your Splunk deployment so that future access is much faster. This current example is atypical but shown for comparison to the SQL command. The source will be renamed orig_source |
TRUNCATE TABLE | TRUNCATE TABLE mytable |
source=mytable | DELETE |
INSERT INTO | INSERT INTO mytable VALUES (value1, value2, value3,....) |
Note: see SELECT INTO. Individual records are not added via the search language, but can be added via the API if need be. |
UNION | SELECT mycolumn FROM mytable1 UNION SELECT mycolumn FROM mytable2 |
source=mytable1 | APPEND [SEARCH source=mytable2] | DEDUP mycolumn |
UNION ALL | SELECT * FROM mytable1 UNION ALL SELECT * FROM mytable2 |
source=mytable1 | APPEND [SEARCH source=mytable2] |
DELETE | DELETE FROM mytable WHERE mycolumn=5 |
source=mytable1 mycolumn=5 | DELETE |
UPDATE | UPDATE mytable SET column1=value, column2=value,... WHERE some_column=some_value |
Note: There are a few things to think about when updating records in . First, you can just add the new values to your Splunk deployment (see INSERT INTO) and not worry about deleting the old values, because Splunk software always returns the most recent results first. Second, on retrieval, you can always de-duplicate the results to ensure only the latest values are used (see SELECT DISTINCT). Finally, you can actually delete the old records (see DELETE). |
See also
Command types | SPL data types and clauses |
This documentation applies to the following versions of Splunk® Enterprise: 7.0.0, 7.0.1, 7.0.2, 7.0.3, 7.0.4, 7.0.5, 7.0.6, 7.0.7, 7.0.8, 7.0.9, 7.0.10, 7.0.11, 7.0.13, 7.1.0, 7.1.1, 7.1.2, 7.1.3, 7.1.4, 7.1.5, 7.1.6, 7.1.7, 7.1.8, 7.1.9, 7.1.10, 7.2.0, 7.2.1, 7.2.2, 7.2.3, 7.2.4, 7.2.5, 7.2.6, 7.2.7, 7.2.8, 7.2.9, 7.2.10, 7.3.0, 7.3.1, 7.3.2, 7.3.3, 7.3.4, 7.3.5, 7.3.6, 7.3.7, 7.3.8, 7.3.9, 8.0.0, 8.0.1, 8.0.2, 8.0.3, 8.0.4, 8.0.5, 8.0.6, 8.0.7, 8.0.8, 8.0.9, 8.0.10, 8.1.0, 8.1.2, 8.1.3, 8.1.4, 8.1.5, 8.1.6, 8.1.7, 8.1.8, 8.1.9, 8.1.10, 8.1.12, 8.2.0, 8.2.1, 8.2.2, 8.2.3, 8.2.4, 8.2.5, 8.2.6, 8.2.7, 8.2.8, 8.2.9, 8.2.10, 8.2.11, 8.2.12, 9.0.0, 9.0.1, 9.0.2, 9.0.3, 9.0.4, 9.0.5, 9.0.6, 9.0.7, 9.0.8, 9.0.9, 9.0.10, 9.1.0, 9.1.1, 9.1.2, 9.1.3, 9.1.4, 9.1.5, 9.1.6, 9.1.7, 9.2.0, 9.2.1, 9.2.2, 9.2.3, 9.2.4, 9.3.0, 9.3.1, 9.3.2, 8.1.1, 8.1.11, 8.1.13, 8.1.14
Feedback submitted, thanks!