ListFile

Description:

Retrieves a listing of files from the input directory. For each file listed, creates a FlowFile that represents the file so that it can be fetched in conjunction with FetchFile. This Processor is designed to run on Primary Node only in a cluster when 'Input Directory Location' is set to 'Remote'. If the primary node changes, the new Primary Node will pick up where the previous node left off without duplicating all the data. When 'Input Directory Location' is 'Local', the 'Execution' mode can be anything, and synchronization won't happen. Unlike GetFile, this Processor does not delete any data from the local filesystem.

Additional Details...

Tags:

file, get, list, ingest, source, filesystem

Properties:

In the list below, the names of required properties appear in bold. Any other properties (not in bold) are considered optional. The table also indicates any default values, and whether a property supports the NiFi Expression Language.

Display NameAPI NameDefault ValueAllowable ValuesDescription
Input DirectoryInput DirectoryThe input directory from which files to pull files
Supports Expression Language: true (will be evaluated using variable registry only)
Listing Strategylisting-strategyTracking Timestamps
  • Tracking Timestamps This strategy tracks the latest timestamp of listed entity to determine new/updated entities. Since it only tracks few timestamps, it can manage listing state efficiently. However, any newly added, or updated entity having timestamp older than the tracked latest timestamp can not be picked by this strategy. For example, such situation can happen in a file system if a file with old timestamp is copied or moved into the target directory without its last modified timestamp being updated. Also may miss files when multiple subdirectories are being written at the same time while listing is running.
  • Tracking Entities This strategy tracks information of all the listed entities within the latest 'Entity Tracking Time Window' to determine new/updated entities. This strategy can pick entities having old timestamp that can be missed with 'Tracking Timestamps'. Works even when multiple subdirectories are being written at the same time while listing is running. However additional DistributedMapCache controller service is required and more JVM heap memory is used. See the description of 'Entity Tracking Time Window' property for further details on how it works.
  • No Tracking This strategy lists an entity without any tracking. The same entity will be listed each time on executing this processor. It is recommended to change the default run schedule value. Any property that related to the persisting state will be disregarded.
Specify how to determine new/updated entities. See each strategy descriptions for detail.
Recurse SubdirectoriesRecurse Subdirectoriestrue
  • true
  • false
Indicates whether to list files from subdirectories of the directory
Record Writerrecord-writerController Service API:
RecordSetWriterFactory
Implementations: FreeFormTextRecordSetWriter
CSVRecordSetWriter
ParquetRecordSetWriter
RecordSetWriterLookup
ScriptedRecordSetWriter
XMLRecordSetWriter
JsonRecordSetWriter
AvroRecordSetWriter
Specifies the Record Writer to use for creating the listing. If not specified, one FlowFile will be created for each entity that is listed. If the Record Writer is specified, all entities will be written to a single FlowFile instead of adding attributes to individual FlowFiles.
Input Directory LocationInput Directory LocationLocal
  • Local Input Directory is located on a local disk. State will be stored locally on each node in the cluster.
  • Remote Input Directory is located on a remote system. State will be stored across the cluster so that the listing can be performed on Primary Node Only and another node can pick up where the last node left off, if the Primary Node changes
Specifies where the Input Directory is located. This is used to determine whether state should be stored locally or across the cluster.
File FilterFile Filter[^\.].*Only files whose names match the given regular expression will be picked up
Path FilterPath FilterWhen Recurse Subdirectories is true, then only subdirectories whose path matches the given regular expression will be scanned
Include File AttributesInclude File Attributestrue
  • true
  • false
Whether or not to include information such as the file's Last Modified Time and Owner as FlowFile Attributes. Depending on the File System being used, gathering this information can be expensive and as a result should be disabled. This is especially true of remote file shares.
Minimum File AgeMinimum File Age0 secThe minimum age that a file must be in order to be pulled; any file younger than this amount of time (according to last modification date) will be ignored
Maximum File AgeMaximum File AgeThe maximum age that a file must be in order to be pulled; any file older than this amount of time (according to last modification date) will be ignored
Minimum File SizeMinimum File Size0 BThe minimum size that a file must be in order to be pulled
Maximum File SizeMaximum File SizeThe maximum size that a file can be in order to be pulled
Ignore Hidden FilesIgnore Hidden Filestrue
  • true
  • false
Indicates whether or not hidden files should be ignored
Target System Timestamp Precisiontarget-system-timestamp-precisionAuto Detect
  • Auto Detect Automatically detect time unit deterministically based on candidate entries timestamp. Please note that this option may take longer to list entities unnecessarily, if none of entries has a precise precision timestamp. E.g. even if a target system supports millis, if all entries only have timestamps without millis, such as '2017-06-16 09:06:34.000', then its precision is determined as 'seconds'.
  • Milliseconds This option provides the minimum latency for an entry from being available to being listed if target system supports millis, if not, use other options.
  • Seconds For a target system that does not have millis precision, but has in seconds.
  • Minutes For a target system that only supports precision in minutes.
Specify timestamp precision at the target system. Since this processor uses timestamp of entities to decide which should be listed, it is crucial to use the right timestamp precision.
Entity Tracking State Cacheet-state-cacheController Service API:
DistributedMapCacheClient
Implementations: HBase_2_ClientMapCacheService
SimpleRedisDistributedMapCacheClientService
HazelcastMapCacheClient
CouchbaseMapCacheClient
RedisDistributedMapCacheClientService
DistributedMapCacheClientService
CassandraDistributedMapCache
Listed entities are stored in the specified cache storage so that this processor can resume listing across NiFi restart or in case of primary node change. 'Tracking Entities' strategy require tracking information of all listed entities within the last 'Tracking Time Window'. To support large number of entities, the strategy uses DistributedMapCache instead of managed state. Cache key format is 'ListedEntities::{processorId}(::{nodeId})'. If it tracks per node listed entities, then the optional '::{nodeId}' part is added to manage state separately. E.g. cluster wide cache key = 'ListedEntities::8dda2321-0164-1000-50fa-3042fe7d6a7b', per node cache key = 'ListedEntities::8dda2321-0164-1000-50fa-3042fe7d6a7b::nifi-node3' The stored cache content is Gzipped JSON string. The cache key will be deleted when target listing configuration is changed. Used by 'Tracking Entities' strategy.
Entity Tracking Time Windowet-time-window3 hoursSpecify how long this processor should track already-listed entities. 'Tracking Entities' strategy can pick any entity whose timestamp is inside the specified time window. For example, if set to '30 minutes', any entity having timestamp in recent 30 minutes will be the listing target when this processor runs. A listed entity is considered 'new/updated' and a FlowFile is emitted if one of following condition meets: 1. does not exist in the already-listed entities, 2. has newer timestamp than the cached entity, 3. has different size than the cached entity. If a cached entity's timestamp becomes older than specified time window, that entity will be removed from the cached already-listed entities. Used by 'Tracking Entities' strategy.
Supports Expression Language: true (will be evaluated using variable registry only)
Entity Tracking Initial Listing Targetet-initial-listing-targetAll Available
  • Tracking Time Window Ignore entities having timestamp older than the specified 'Tracking Time Window' at the initial listing activity.
  • All Available Regardless of entities timestamp, all existing entities will be listed at the initial listing activity.
Specify how initial listing should be handled. Used by 'Tracking Entities' strategy.
Entity Tracking Node Identifieret-node-identifier${hostname()}The configured value will be appended to the cache key so that listing state can be tracked per NiFi node rather than cluster wide when tracking state is scoped to LOCAL. Used by 'Tracking Entities' strategy.
Supports Expression Language: true (will be evaluated using variable registry only)
Track Performancetrack-performancefalse
  • true
  • false
Whether or not the Processor should track the performance of disk access operations. If true, all accesses to disk will be recorded, including the file being accessed, the information being obtained, and how long it takes. This is then logged periodically at a DEBUG level. While the amount of data will be capped, this option may still consume a significant amount of heap (controlled by the 'Maximum Number of Files to Track' property), but it can be very useful for troubleshooting purposes if performance is poor is degraded.
Maximum Number of Files to Trackmax-performance-metrics100000If the 'Track Performance' property is set to 'true', this property indicates the maximum number of files whose performance metrics should be held onto. A smaller value for this property will result in less heap utilization, while a larger value may provide more accurate insights into how the disk access operations are performing
Supports Expression Language: true (will be evaluated using variable registry only)
Max Disk Operation Timemax-operation-time10 secsThe maximum amount of time that any single disk operation is expected to take. If any disk operation takes longer than this amount of time, a warning bulletin will be generated for each operation that exceeds this amount of time.
Supports Expression Language: true (will be evaluated using variable registry only)
Max Directory Listing Timemax-listing-time3 minsThe maximum amount of time that listing any single directory is expected to take. If the listing for the directory specified by the 'Input Directory' property, or the listing of any subdirectory (if 'Recurse' is set to true) takes longer than this amount of time, a warning bulletin will be generated for each directory listing that exceeds this amount of time.
Supports Expression Language: true (will be evaluated using variable registry only)

Relationships:

NameDescription
successAll FlowFiles that are received are routed to success

Reads Attributes:

None specified.

Writes Attributes:

NameDescription
filenameThe name of the file that was read from filesystem.
pathThe path is set to the relative path of the file's directory on filesystem compared to the Input Directory property. For example, if Input Directory is set to /tmp, then files picked up from /tmp will have the path attribute set to "/". If the Recurse Subdirectories property is set to true and a file is picked up from /tmp/abc/1/2/3, then the path attribute will be set to "abc/1/2/3/".
absolute.pathThe absolute.path is set to the absolute path of the file's directory on filesystem. For example, if the Input Directory property is set to /tmp, then files picked up from /tmp will have the path attribute set to "/tmp/". If the Recurse Subdirectories property is set to true and a file is picked up from /tmp/abc/1/2/3, then the path attribute will be set to "/tmp/abc/1/2/3/".
file.ownerThe user that owns the file in filesystem
file.groupThe group that owns the file in filesystem
file.sizeThe number of bytes in the file in filesystem
file.permissionsThe permissions for the file in filesystem. This is formatted as 3 characters for the owner, 3 for the group, and 3 for other users. For example rw-rw-r--
file.lastModifiedTimeThe timestamp of when the file in filesystem was last modified as 'yyyy-MM-dd'T'HH:mm:ssZ'
file.lastAccessTimeThe timestamp of when the file in filesystem was last accessed as 'yyyy-MM-dd'T'HH:mm:ssZ'
file.creationTimeThe timestamp of when the file in filesystem was created as 'yyyy-MM-dd'T'HH:mm:ssZ'

State management:

ScopeDescription
LOCAL, CLUSTERAfter performing a listing of files, the timestamp of the newest file is stored. This allows the Processor to list only files that have been added or modified after this date the next time that the Processor is run. Whether the state is stored with a Local or Cluster scope depends on the value of the <Input Directory Location> property.

Restricted:

This component is not restricted.

Input requirement:

This component does not allow an incoming relationship.

System Resource Considerations:

None specified.

See Also:

GetFile, PutFile, FetchFile