brianmhess/cassandra-loader

Introductioncassandra-loader is a general-purpose, delimited-file, bulk loader forCassandra. It supports a number of configuration options to enable bulkloading of various types of delimited files, includingcomma-separated valuestab-separated valuescustomer delimiter-separated valuesheader rowcomma as decimal separator...Getting itDownloadingThis utility has already been built, and is available athttps://github.com/brianmhess/cassandra-loader/releases/download/v0.0.27/cassandra-loaderGet it with wget:wget https://github.com/brianmhess/cassandra-loader/releases/download/v0.0.27/cassandra-loaderBuildingTo build this repository, simply clone this repo and run:gradle loaderAll of the dependencies are included (namely, the Java driver - currentlyversion 3.0.0). The output will be the cassandra-loader executablein the build directory. There will also be an jar with all of thedependencies included in the build/libs/cassandra-loader-uber-.jarDocumentationTo extract this README document, simply run (on the cassandra-loaderexecutable - (e.g., on build/cassandra-loader):jar xf cassandra-loader README.mdRunTo run cassandra-loader, simply run the cassandra-loader executable(e.g., located at build/cassandra-loader):cassandra-loaderIf you built this with gradle, you can also run:gradle runThis will print the usage statement.The following will load the myFileToLoad.csv file into the Cassandracluster at IP address 1.2.3.4 into the test.ltest column family wherethe myFileToLoad file has the format of 4 columns - and it gets thedata type information from the database - and using the default options:cassandra-loader -f myFileToLoad.csv -host 1.2.3.4 -schema "test.ltest(a, b, c, d)"Options:SwitchOptionDefaultDescription-configFileFilenamenoneFilename of configuration options-fFilename<REQUIRED>Filename to load - required.-hostIP Address<REQUIRED>Cassandra connection point - required.-formatInput formatdelimFormat of the data. Options are "delim" or "json".-schemaCQL schemaSchema of input data - required for delim In the format "keySpace.table(col1,col2,...)" and in the order that the data will be in the file.-keyspaceKeyspace nameName of keyspace (case sensitive) to load in to - required for json-tableTable nameName of table (case sensitive) to load in to - required for json-portPort Number9042Cassandra native protocol port number-userUsernamenoneCassandra username-pwPasswordnoneCassandra password-ssl-truststore-pathTruststore PathnonePath to SSL truststore-ssl-truststore-pwTruststore PasswordnonePassword to SSL truststore-ssl-keystore-pathKeystore PathnonePath to SSL keystore-ssl-keystore-pathKeystore PasswordnonePassword to SSL keystore-consistencyLevelConsistency LevelONECQL Consistency Level-numThreadsNumber of threadsNumber of CPUsNumber of threads to use (one per file)-numFuturesNumber of Futures1000Number of Java driver futures in flight.-numRetriesNumber of retries1Number of times to retry the INSERT before declaring defeat.-queryTimeoutTimeout in seconds2Amount of time to wait for a query to finish before timing out.-ttlTime To LivenoneTTL to use when inserting these rows-delimDelimiter,Delimiter to use-charsPerColumnCharacters per column4096Maximum characters per column-nullStringNull String<empty string>String to represent NULL data-boolStyleBoolean StyleTRUE_FALSEString for boolean values. Options are "1_0", "Y_N", "T_F", "YES_NO", "TRUE_FALSE".-decimalDelimDecimal delimiter.Delimiter for decimal values. Options are "." or ","-dateFormatDate Format Stringdefault for Locale.ENGLISHDate format string as specified in the SimpleDateFormat Java class: http://docs.oracle.com/javase/7/docs/api/java/text/SimpleDateFormat.html-skipRowsRows to skip0Number of rows to skip at the beginning of the file-skipColsColumns to skip<not set>Comma-separated list of columns to skip loading (0-counted)-maxRowsMax rows to read-1Maximum rows to read (after optional skipping of rows). -1 signifies all rows.-maxErrorsMax parse errors10Maximum number of rows that do not parse to allow before exiting.-maxInsertErrorsMax insert errors10Maximum number of rows that do not insert to allow before exiting.-badDirBad directorycurrent directoryDirectory to write badly parsed and badly inserted rows - as well as the log file.-rateIngest rateunlimitedMaximum rate to insert data - in rows/sec.-progressRateProgress rate100000How often to report the ingest rate (number of rows)-rateFileRate Stats File<not set>File to contain CSV rate statistics-successDirSuccess directory<not set>Location to move successfully loaded files-failureDirFailure directory<not set>Location to move files that failed to load-batchSizeBatch size1Size of unlogged batches. If set to 1 then no batching.-commentComment character$lt;not set>Comment character.CommentsUsing stdinYou can send data in on stdin by specifying the filename (via the -f switch) as "stdin" (case insensitive).That way, you could pipe data in from other commands:grep IMPORTANT data.csv | cassandra-loader -f stdin -h 1.2.3.4 -schema "test.itest(a, b)"Support for collectionsCollections are supported. Their format is the CQL native one.Sets are started with '{' and ended with '}' and enclose a comma-separated list{1,2,3} or {"a","b","c"}Lists are started with '[' and ended with ']' and enclose a comma-separated list[1,2,3] or ["a","b","c"]Maps are started with '{' and ended with '}' and enclose a comma-separated listof pairs that are separated by ':'{1:1,2:2,3:3} or {"a":1, "b":2, "c":3}All collections must be enclosed in double-quotes.Username/PasswordIf you specify either the username or the password, then you must specify both.Boolean StyleboolStyle is a case-insensitive test of the True and False strings. For thedifferent styles, the True and False strings are as follows: Style | True | False------------|------|------- 0_1 | 1 | 0 Y_N | Y | N T_F | T | F YES_NO | YES | NO TRUE_FALSE | TRUE | FALSE Configuration fileconfigFile is a file with configuration options that are formatted just like onthe command line. This allows you to not specify arguments on the command line.For example, you can specify passwords in the configFile and avoid having them onthe command line. The format is one switch and option per line:-pwd mypassword-host 1.2.3.4MiscelaneousnumFutures is a way to control the level of parallelism, but at some pointtoo many will actually slow down the load. The default of 500 is a decentplace to start.If you use quoted strings, you need to use double-quotes. To escape a double-quote inside a quoted string, use the backslash to escape it ("""). To create a backslash inside a quoted string, use two backslashes in a row ("\"). If you quote your string, it will not be trimmed, but if you do not quote your string it will be trimmed.Loading into counter columns is not supported.The default nullString is the empty string. If you want empty strings to be saved as empty strings, set the nullString to something else.If you do not set the successDir then files that successfully loaded will remain in their input directory. The same is true for failed files if you do not set the failureDir. You cannot set either if the input file is "stdin".When using jsonline, all JSON field names are case-sensitive. When using jsonline or jsonarray, the -keyspace and -table arguments are case-sensitive.Usage Statement:version: 0.0.27Usage: -f <filename> -host <ipaddress> [OPTIONS]OPTIONS: -schema <schema> Table schema (when using delim) -table <tableName> Table name (when using json) -keyspace <keyspaceName> Keyspace name (when using json) -configFile <filename> File with configuration options -delim <delimiter> Delimiter to use [,] -charsPerColumn <chars> Max number of chars per column [4096] -dateFormat <dateFormatString> Date format for TIMESTAMP [default for Locale.ENGLISH] -localDateFormat <formatString> Date format for DATE [yyyy-MM-dd] -nullString <nullString> String that signifies NULL [none] -comment <commentString> Comment symbol to use [none] -skipRows <skipRows> Number of rows to skip [0] -skipCols <columnsToSkip> Comma-separated list of columsn to skip in the input file -maxRows <maxRows> Maximum number of rows to read (-1 means all) [-1] -maxErrors <maxErrors> Maximum parse errors to endure [10] -badDir <badDirectory> Directory for where to place badly parsed rows. [none] -port <portNumber> CQL Port Number [9042] -user <username> Cassandra username [none] -pw <password> Password for user [none] -ssl-truststore-path <path> Path to SSL truststore [none] -ssl-truststore-pw <pwd> Password for SSL truststore [none] -ssl-keystore-path <path> Path to SSL keystore [none] -ssl-keystore-pw <pwd> Password for SSL keystore [none] -consistencyLevel <CL> Consistency level [LOCAL_ONE] -numFutures <numFutures> Number of CQL futures to keep in flight [1000] -batchSize <batchSize> Number of INSERTs to batch together [1] -decimalDelim <decimalDelim> Decimal delimiter [.] Other option is ',' -boolStyle <boolStyleString> Style for booleans [TRUE_FALSE] -numThreads <numThreads> Number of concurrent threads (files) to load [num cores] -queryTimeout <# seconds> Query timeout (in seconds) [2] -numRetries <numRetries> Number of times to retry the INSERT [1] -maxInsertErrors <# errors> Maximum INSERT errors to endure [10] -rate <rows-per-second> Maximum insert rate [50000] -progressRate <num txns> How often to report the insert rate [100000] -rateFile <filename> Where to print the rate statistics -successDir <dir> Directory where to move successfully loaded files -failureDir <dir> Directory where to move files that did not successfully load -nullsUnset [false|true] Treat nulls as unset [faslse] -format [delim|jsonline|jsonarray] Format of data: delimited or JSON [delim] -table <tableName> Table name (when using JSON) -keyspace <keyspaceName> Keyspace name (when using JSON) -ttl <TTL> TTL for all rows in this invocation [unset]Examples:cassandra-loader -f /path/to/file.csv -host localhost -schema "test.test3(a, b, c)"cassandra-loader -f /path/to/directory -host 1.2.3.4 -schema "test.test3(a, b, c)" -delim "\t" -numThreads 10cassandra-loader -f stdin -host localhost -schema "test.test3(a, b, c)" -user myuser -pw mypassword##Examples:Load file /path/to/file.csv into the test3 table in the test keyspace usingthe cluster at localhost. Use the default options:cassandra-loader -f /path/to/file.csv -host localhost -schema "test.test3(a, b, c)"Load all the files from /path/to/directory into the test3 table in the testkeyspace using the cluster at 1.2.3.4. Use 10 threads and use tab as thedelimiter:cassandra-loader -f /path/to/directory -host 1.2.3.4 -schema "test.test3(a, b, c)" -delim "\t" -numThreads 10Load the data from stdin into the test3 table in the test keyspace using thecluster at localhost. Use "myuser" as the username and "mypassword" as thepassword:cassandra-loader -f stdin -host localhost -schema "test.test3(a, b, c)" -user myuser -pw mypassword##SampleIncluded here is a set of sample data. It is in the sample/ directory.You can set up the table and keyspace by running:cqlsh -f sample/cassandra-schema.cqlTo load the data, run:cd sample./load.shTo check that things have succeeded, you can run:wc -l titanic.csvAnd:cqlsh -e "SELECT COUNT(*) FROM titanic.surviors"Both should return 891.cassandra-unloadercassandra-unloader is a utility to dump the contentsof a Cassandra table to delimited file format. It usesthe same sorts of options as cassandra-loader so that theoutput of cassandra-unloader could be piped intocassandra-loader:cassandra-unloader -f stdout -host host1 -schema "ks.table(a,b,c)" | cassandra-loader -f stdin -host host2 -schema "ks2.table2(x,y,z)"Get it with wget:wget https://github.com/brianmhess/cassandra-loader/releases/download/v0.0.27/cassandra-unloaderTo build, run:gradle unloaderTo run cassandra-unloader, simply run the cassandra-unloader executable(e.g., located at build/cassandra-unloader):cassandra-unloader###Usage statement:version: 0.0.27Usage: -f <outputStem> -host <ipaddress> -schema <schema> [OPTIONS]OPTIONS: -configFile <filename> File with configuration options -format [delim|jsonline|jsonarray] Format of data: delimited or JSON [delim] -delim <delimiter> Delimiter to use [,] -dateFormat <dateFormatString> Date format for TIMESTAMP [default for Locale.ENGLISH] -localDateFormat <FormatString> Date format for DATE [yyyy-MM-dd] -nullString <nullString> String that signifies NULL [none] -port <portNumber> CQL Port Number [9042] -user <username> Cassandra username [none] -pw <password> Password for user [none] -ssl-truststore-path <path> Path to SSL truststore [none] -ssl-truststore-pw <pwd> Password for SSL truststore [none] -ssl-keystore-path <path> Path to SSL keystore [none] -ssl-keystore-pw <pwd> Password for SSL keystore [none] -consistencyLevel <CL> Consistency level [LOCAL_ONE] -decimalDelim <decimalDelim> Decimal delimiter [.] Other option is ',' -boolStyle <boolStyleString> Style for booleans [TRUE_FALSE] -numThreads <numThreads> Number of concurrent threads to unload [5] -beginToken <tokenString> Begin token [none] -endToken <tokenString> End token [none] -where <predicate> WHERE clause [none] -fetchSize <fetchSize> Fetch size to use [0]A few simple examples using the -where are as follows:cassandra-unloader -host localhost -f stdout -schema "testks.testtable(pkey,ccol,x,y)" -where "pkey=5"cassandra-unloader -host localhost -f stdout -schema "testks.testtable(pkey,ccol,x,y)" -where "x = 100 ALLOW FILTERING"