Shareware Grab Bag

home *** CD-ROM | disk | FTP | other *** search

/ Shareware Grab Bag / Shareware Grab Bag.iso / 090 / sampler4.arc / IO.SAC < prev next >

Wrap

Text File | 1985-08-30 | 28.6 KB | 841 lines

ON ECHO !*******************************************************************! ! ! ! Data Input and Output in SORITEC Sampler ! ! (Chapters 3 and 4) ! ! ! !*******************************************************************! ! ! SORITEC Sampler can import and export data in a number of formats. ! In addition, Sampler supports a complete databanking facility that ! allows you to maintain all of SORITEC's data types, plus GROUPS and ! equations with a simple set of commands. This demonstration ! illustrates most of these capabilities. ! !*******************************************************************! ! ! ! Keyboard Entry ! ! (Section 3.4) ! ! ! !*******************************************************************! ! ! Data are entered directly from the terminal keyboard with the ! FILL command, e.g., ! FILL short_series_a 4 6 8 3 6 8 3 5 8 2 ! ! If no USE command is in effect, SORITEC Sampler counts the number of ! observations associated with the variable and sets the USE period ! accordingly. Let's enter the USE command with no arguments to ! see how Sampler defined the USE period for the previous FILL command. ! USE ! ! You can use a standard FORTRAN repeat operator to enter consecutive ! numbers with the same value, e.g., ! FILL short_series_b 5 3*6 4 9.3 2*10.1 8 14 ! ! If you enter more observations for a variable than defined by ! the current USE period, SORITEC Sampler returns an error and ! ignores the command, e.g., ! FILL too_long 3 6 8 4 6 7 3 6 9 5 3 8 ! ! You must change the USE period to accommodate the number of ! observations when there are more data to be entered than allowed ! by the current USE period. Any periodicity of time series, ! included undated data types, may be entered from the terminal. ! USE 1980m1 1980m12 FILL too_long 3 6 8 4 6 7 3 6 9 5 3 8 ! ! Normally, Sampler generates an error and ignores the command when ! too few data items are entered. ! FILL too_short 4 5 7 3 4 ! ! However, if you enable the RAGGED option with the ON RAGGED command, ! data series shorter than the current USE period are accepted. The ! remaining observations are assigned MISSING values. ! ON RAGGED FILL too_short 4 5 7 3 4 OFF RAGGED PRINT too_long too_short ! ! The FILL command is commonly used to enter short data series or to ! extend or revise existing data series. For example, extending the ! variable "too_short" be three observations is easy. ! USE 1980m6 1980m8 FILL new_data 4 5 8 REVISE too_short = new_data USE 1980m1 1980m12 PRINT too_long too_short ! ! More information about data revision in SORITEC Sampler is given ! in the file, "revise.sac". ! !*******************************************************************! ! ! ! SAL File Input and Output ! ! (Section 3.1) ! ! ! !*******************************************************************! ! ! SORITEC SAL files are the most convenient way to transfer data into ! Sampler. SAL files are simply unformatted flat-ASCII files that ! have special headers and trailers that identify the USE period ! and variable names. A sample SAL file is shown below. ! ! USE 1978 1983 ! READ total_population ! 222585.0000 225055.0000 227738.0000 230019.0000 ! 232309.0000 234496.0000 ! ; ! READ male_population ! 108424.0000 109584.0000 110874.0000 111984.0000 ! 113105.0000 114179.0000 ! ; ! END ! READ female_population ! 114161.0000 115472.0000 116864.0000 118034.0000 ! 119204.0000 120317.0000 ! ; ! END ! ! To read a SAL file, simply enter READ(filename). The header ! informs Sampler what the USE period of the data is and what ! the variable names are. The file above is stored under the name ! "popdata.sal" on this demonstration diskette, which is assumed ! to be in the a: drive. To read it, simply enter: ! READ('a:popdata') ! ! There are several things to note about this command: ! (1) The .SAL extension is omitted from the filename argument. ! Any file that is read with the READ command must have a ! .SAL extension. Including the file extension will generate ! an error. ! (2) If the file is in other than the current directory, the ! filename must be enclosed within single quotes. ! (3) In executing the READ statement, SORITEC Sampler reads ! data until it encounters an END statement. A subsequent ! READ of the same file continues reading the file until ! another END statement is found. ! ! In the example above, SORITEC Sampler has read "total_population" ! and "male_population" into the workspace. You can find this out ! by entering the command SYMBOLS, e.g., ! SYMBOLS ! ! Let's READ in "female_population" now with another READ command. ! READ('a:popdata') SYMBOLS ! ! A SAL file can contain as many blocks, delimited by END statements ! as disk space permits. Also, SAL files can contain variables of ! mixed periodicities. Simply insert a USE statement before the READ ! statement that identifies the variable. ! ! Data may be written to files in SAL file format with the PUNCH command. ! This is particularly useful if you wish to export data to another ! SORITEC on another type of computer since data transfer ! is essentially transparent and yet the data are in ASCII format. ! ! PUNCH only writes to a file named PUNCH1.SAL in the current directory. ! Unlike file naming in the READ command, you cannot direct output to any ! other file name. Use the DOS REN command to change the file name after ! terminating SORITEC Sampler. Note that if a file PUNCH1.SAL already ! exists on your directory, SORITEC Sampler will delete it and open a new ! file PUNCH1.SAL. The PUNCH command will be superceded by more general ! output commands in later versions of SORITEC. ! ! Let's say we want to write a SAL file containing (1) the difference ! between female and male populations, and (2) the ratio of female to ! total population. First, make the appropriate calculations. ! pop_difference = female_population - male_population female_proportion = female_population/total_population ! ! Now write out a SAL file called to the current directory. If you ! are running Sampler off a hard disk system, the file will be written ! to the directory which contains the Sampler program. If you are ! running the program on a floppy disk system, the file will be written ! on the a: drive. We'll only demonstrate the PUNCH command with a ! small amount of data since space is limited if you are running Sampler ! on a double-floppy system. ! PUNCH pop_difference female_proportion ! ! Note that SORITEC Sampler leaves the SAL file open for ! additional data output until you PUNCH to another file or the ! SORITEC Sampler is terminated. When the file is closed, an END ! statement is appended to it. No intervening END statements are ! inserted into the file, for example, after each PUNCH statement ! is executed. Remember, if PUNCH1.SAL already exists on the current ! directory, it is deleted before the data associated with the PUNCH ! command are written. Unlike SORITEC Databanks, SAL files can only ! be "managed" by DOS commands. ! !*******************************************************************! ! ! ! DIF File Input and Output ! ! (Section 3.2) ! ! ! !*******************************************************************! ! ! Data can be transferred between most population PC packages that ! support the DIF file format. ! ! For DIF file input, use the READDIF command. Although the command syntax ! is similar to the READ command, there are some important differences. ! (1) READDIF does not interpret dates in the DIF file so you ! must specify the USE period for the data before you enter ! the READDIF command. ! (2) If variable names are not in the DIF file, they must be ! entered in the command line. ! ! As an example, consider the DIF file generated by LOTUS' Translate ! utility. If you have a copy of LOTUS, this worksheet from which the ! data were translated is in "POPNEW.WKS". The worksheet consists of ! two labelled columns containing the variables "popciv" and "poh" for ! the period 1982m1 to 1983m12. The worksheet has been translated to a ! file named "popnew.dif". ! ! To import the data, set the USE period: ! USE 1982m1 1983m12 ! ! then read the DIF file with the command: ! READDIF('a:popnew') ! ! If the DIF file was not labelled, the command line would be: ! ! READDIF('a:popnew') popciv poh ! ! ! Data are exported in DIF file format using the PUNCHDIF command. ! For example, if you wanted to export the 1982 values of "poh" and ! "popciv" to the file "popdata.dif", the command would be: ! USE 1982m1 1982m12 PUNCHDIF('a:popdata') poh popciv ! ! If you have a copy of LOTUS, use the Translate utility to convert ! the file into a worksheet. Note that LOTUS translates the dates into ! a variable called "time" and writes these data to the first column ! (or row) of the worksheet. When dates are in the DIF file, SORITEC ! Sampler assigns MISSING values to "time". DIF files are managed by ! DOS, so use DOS commands to delete, copy, rename, etc. the files. ! !*******************************************************************! ! ! ! Formatted Input and Output ! ! (Section 3.2) ! ! ! !*******************************************************************! ! ! Formatted input and output is similar to that in FORTRAN programs. ! The format of the data that are to be imported or exported are ! specified in a FORMAT statement, identified by a statement number. ! Data are read or written in the desired format by refering to the ! statement number that defines that format. Data may be read or ! written in formatted form either to files or the terminal with the ! READ and WRITE commands. ! ! Note that you may WRITE data to an output device without refering ! to a FORMAT statment number. In this case, data are written as ! they are with the PRINT command, e.g., writing "male_population" ! and "female_population" to the terminal results in the following ! output: ! USE 1978 1983 WRITE male_population female_population ! ! Formatted input and output is demonstrated in an example whereby ! we wish to read from a file data in tabular form. The data are ! defined as in the following table. ! ! 1st row --> KEY MACROECONOMIC INDICATORS ! ! Year Gross Nat. Product Civ. Labor Force Corporate Profits ! (billions of $) (millions of persons) (billions of $) ! 1980 2631.7 106.940 149.829 ! 1981 2957.8 108.670 140.009 ! 1983 3069.3 110.205 104.843 ! 1984 3304.8 111.551 127.414 ! ! ^ ! | ! 1st column ! ! You should always leave a blank line at the end of the file to avoid ! spurious READ errors. ! ! To read the data, first set up the format statement. Use of integer ! (I) format is not advised for input so the year data should be read ! in as REAL. ! 100 FORMAT(///4(/f4.0,8x,f6.1,15x,f6.0,15x,f6.0)) ! ! Next, define the USE period for the data. ! USE 1980 1983 ! ! When importing data with a formatted READ, SORITEC Sampler normally ! expects to read in one observation at a time, re-using the format ! from the start for each observation. If you are putting more than ! one observation on each input record, or you need to skip some initial ! input records, you need to enable the STREAMIO global option using ! the ON STREAMIO command. This causes SORITEC to compute the total ! number of data points to be read (i.e., the number of observations ! multiplied by the number of variables), and then read the entire ! mass of data in a single formatted READ, reusing the format only as ! and when necessary. ! ON STREAMIO ! ! Now import the data with the READ command. We'll read the data ! directly from this command file to save space on the demonstration ! diskette. ! READ(100) year gnp clf profits KEY MACROECONOMIC INDICATORS Year Gross Nat. Product Civ. Labor Force Corporate Profits (billions of $) (millions of persons) (billions of $) 1980 2631.7 106940 149829 1981 2957.8 108670 140009 1983 3069.3 110205 104843 1984 3304.8 111551 127414 ! ! Print out the values of the variables for verification. ! PRINT year gnp clf profits ! ! Let's add population to the table and print it out row-wise rather then ! column-wise. We'll write the table to the terminal instead of disk to ! save some room. ! ! Set up the format statement. ! 200 FORMAT(//10x,'Key Macroeconomic Indicators Including Population'/) 201 FORMAT(19x,4(8x,f5.0)/18x,4(9x,4('-'))) 202 FORMAT(1x,'Gross Nat. Product',4(3x,f10.1)/' (billions of $)'/) 203 FORMAT(1x,'Civ. Labor Force ',4(5x,f8.0)/' (thousands of persons)'/) 204 FORMAT(1x,'Corporate Profits ',4(5x,f8.0)/' (millions of $)'/) 205 FORMAT(1x,'Total Population ',4(5x,f8.0)/' (thousands of persons)'/) ! ! Use the WRITE command to write out the table. Note that to prevent ! the echoing of the WRITE commands in between the output, disable the ! ECHO option with the OFF ECHO command. ! OFF ECHO WRITE(200) WRITE(201) year WRITE(202) gnp WRITE(203) clf WRITE(204) profits WRITE(205) total_population ON ECHO ! ! If this table had been written to disk, it would have been ! written with the command: ! ! WRITE(filename 200) ! WRITE(filename 201) year ! WRITE(filename 202) gnp ! WRITE(filename 203) clf ! WRITE(filename 204) profits ! WRITE(filename 205) total_population ! ! The table would reside in the file "filename.lst". ! ! Formatted data can be written to the terminal or file in column-wise ! organization. To do this, enclose the variable list in parentheses. ! For example, if you wanted to write "year", "gnp", "clf" and "profits" ! in the same format as the table that was input, the command would be: ! ! WRITE(filename 100) (year gnp clf profits) ! ! ! !*******************************************************************! ! ! ! SORITEC Sampler Databanking Commands ! ! (Chapter 4) ! ! ! !*******************************************************************! ! ! SORITEC databanks are the key to using Sampler efficiently. Databanks ! can store data series, equations, matrices, vectors, parameters, ! GROUPS and even entire models in the full version of SORITEC. ! SORITEC databanks are created, managed and deleted entirely within ! SORITEC Sampler. In your DOS directory, they are identified by ! the extension, "SDB". This section of the demonstration illustrates ! the capabilities of the SORITEC Sampler's databanking facility. ! ! Data banks are created with the CREATE command (but first we'll ! attempt to PURGE it in case it already exists), e.g., ! PURGE 'a:ecdata' ! ! Don't worry about the message if the file was not found. ! CREATE 'a:ecdata' ! ! Note that when redirecting output to another drive or directory, ! the filename must be enclosed within single quotes. Otherwise, ! simply enter the filename. In out example, we should now have a ! databank file named "ecdata.sdb" on the a: drive. Note that ! the "SDB" extension is not included with the file name in the ! command line. Databank names must be 8 characters or less to conform ! to DOS naming conventions. ! ! You save data in a databank with the KEEP command. For example, ! to save "gnp", "clf", and "profits" data that we used earlier, ! enter: ! KEEP gnp clf profits ! ! To verify the contents of the databank you have currently attached, ! enter the CONTENTS command, e.g., ! CONTENTS ! ! KEEP saves all observations associated your data series regardless ! of the currently active USE period. If you want to save only the ! active observations, you must enter the keyword ACTIVE as a ! modifier to the command. For example, suppose we want to save ! only the 1982 and 1983 values of "pop_difference" and ! "female_proportion" in "ecdata". Then change the USE period to ! the 1982 to 1983 period, and then KEEP them with the ACTIVE ! modifier included in the command line. ! USE 1982 1983 KEEP(ACTIVE) pop_difference female_proportion ! ! Only the observations for 1982 and 1983 will be saved. ! ! Note that once we have these variables in the databank, we can ! remove them from SORITEC Sampler's symbol table with the FORGET ! command, e.g., ! FORGET pop_difference female_proportion ! ! If you type SYMBOLS, you will note that these two variables are not ! longer in Sampler's workspace. While we have them in the "ecdata" ! databank, we only have the observations for 1982 and 1983. Previous ! observations are lost, once cleared from the workspace. ! ! You can explicitly detach a SORITEC databank from your job session ! with the RETURN command. For example, ! RETURN ! ! closes "ecdata.sdb" and detaches it from SORITEC Sampler. Databanks ! are automatically returned when you CREATE a new databank or access ! an existing one. ! ! If you want to attach an existing databank, you ACCESS it, e.g., ! ACCESS 'a:ecdata' ! ! You can implicitly access a databank with the CONTENTS command. For ! example, "CONTENTS ecdata" would return any attached databank, access ! the named databank and produce a listing of variables it contains. ! ! Let's demonstrate some the the things that can be stored in SORITEC ! Sampler's databanks. To do this, we'll create a databank on the "a:" ! drive to store population-related information. Again, we'll attemp ! to PURGE it in case it already exists. ! PURGE 'a:popdata' ! ! If the file does not already exist, you will receive a message. ! CREATE 'a:popdata' ! ! Let's define a GROUP to reference the three annual population variables ! that we started out with in this demonstration, "total_population", ! "male_population" and "female_population". ! GROUP pop_group total_population male_population female_population ! ! Print out the data after restoring the original USE period. ! USE 1978 1983 ON GROUP PRINT pop_group ! ! We can use the GROUP definition to store all population data in a ! single command with one argument, e.g., ! KEEP pop_group ! ! Now if you think you might want to refer to these three population ! variables as a group later, store the GROUP definition, as well, ! by first turning OFF the group expansion option and then KEEPing ! the GROUP. ! OFF GROUP KEEP pop_group ! ! The next time you want to go into this databank, simply retrieve the ! GROUP with the group expansion option turned off, enable group expansion ! with the ON GROUP command and then copy the data from the databank by ! refering to the GROUP definition. ! ! SORITEC Sampler can also store equations. Since we lost some ! observations for "pop_difference" and "female_proportion" because ! we chose only to store the ACTIVE observations, let's define ! equations for the two variables and store the equations in ! the "popdata.sdb" databank so that we can recreate the data from ! the other variables in the databank. ! ! First, define the equations: ! equation pop_dif pop_difference = female_population - male_population equation fem_prop female_proportion = female_population/total_population ! ! and then KEEP them. ! KEEP pop_dif fem_prop ! ! You can similarly store vectors, parameters, constants and matrices ! generated by Sampler. Remember that since Sampler does not support ! matrix operations generally, the only matrices you'll be able to ! store are those generated by the various estimation commands and ! the XTAB command. ! ! Let's forget everything from Sampler's workspace since all the ! important information is now in the two databanks. ! FORGET * ! ! Several other utilities are available to manage items in your ! databanks. To show these let's ACCESS "ecdata.sdb" from the a:drive. ! ACCESS 'a:ecdata' CONTENTS ! ! Suppose we want to rename items in a databank. This is done with the ! RENAME command. For example, to rename "gnp" to "gnp_in_current_dos" ! the command would be: ! USE 1980 1983 RENAME gnp_in_current_dollars gnp ! ! Note the argument order in this command: RENAME new_name old_name ! ! For practice, we'll rename the other two economic items, clf and ! profits so that their units are identified with the variable name. ! RENAME profits_mil_of_dols clf RENAME clf_thous_pers profits ! ! Whoops! We assigned the wrong names to the two variables. Oh well, ! this is a good time to demonstrate the SWITCH command, which allows ! you to switch the names of two variables. ! SWITCH profits_mil_of_dols clf_thous_pers ! ! Now the correct names will be associated with the proper data. ! One important thing to remember about renaming or switching the ! names of items in databanks. If you change the name of an item ! and that item is also defined in a GROUP, the group definition ! is NOT automatically updated and you will be unable to reference ! the data from the original group definition. ! ! ! Retrieving items from a databank is even easier than saving them. ! You can either COPY them explicitly or simply reference them in ! a legal SORITEC Sampler command. If Sampler cannot find the item ! in its workspace it automatically searches any attached databank ! to see if it exists there. If it does, the item is copied to the ! workspace and used as required. ! ! The copy command is simply: ! COPY profits_mil_of_dols clf_thous_pers ! ! Let's make sure the item names are associated with the right data. ! PRINT profits_mil_of_dols clf_thous_pers ! ! Note that these items are now back in Sampler's workspace. ! SYMBOLS ! ! We could have just as easily printed the data without explicitly ! copying it as shown for gnp_in_current_dollars ! PRINT gnp_in_current_dollars ! ! And notice that it, too, is now back in the workspace. ! SYMBOLS ! ! If you want to delete items from an attached databank, use the DISCARD ! command. For example, suppose we want to store "pop_difference" and ! "female_proportion" in the "popdata.sdb" databank. Let's first copy ! the values for "pop_difference" and "female_proportion" into the ! workspace for future use. ! COPY pop_difference female_proportion ! ! Then DISCARD the unwanted items. ! DISCARD pop_difference female_proportion ! ! To check, type CONTENTS, and Voila! ! CONTENTS ! ! We'll access "popdata.sdb" now to complete the transfer of information. ! ACCESS 'a:popdata' ! ! Note that "ecdata.sdb" will be automatically RETURNed with this command. ! ! Now KEEP the two population variables. ! KEEP pop_difference female_proportion ! ! But wait, there are only data for 1982 and 1983. ! PRINT pop_difference female_proportion ! ! We should re-generate the 1978 to 1981 values with the equations ! we stored earlier. Let's check to see what we have stored ! in the databank. ! CONTENTS ! ! We'll retrieve the two equations and the GROUP definition, but will ! let Sampler implicitly copy the data series when they are referenced ! in equation computations. ! COPY pop_dif fem_prop pop_group ! ! Equations and group definitions are printible, so let's take a look. ! PRINT pop_dif fem_prop pop_group ! ! Check the USE period to make sure we are computing over the appropriate ! period. Note that SORITEC Sampler makes no assumptions about the USE ! period when it copies an item from a databank. For data series and any ! other multiple element data type that is stored, Sampler copies all ! observations regardless of active USE period. If no USE period is ! defined, it will still copy all observations, but you will not be able ! to do anything with the data, such as PRINT it, etc. until a valid ! USE period is defined. ! USE ! ! We'll change the period to 1978 to 1981 and REVISE the two series so ! they incorporate all four observations. ! USE 1978 1981 ON REVISE ! ! Equations and identities are calculated with the COMPUTE command. ! COMPUTE pop_dif COMPUTE fem_prop ! ! Disable the data revision option. ! OFF REVISE ! ! Print out the values for 1978 to 1983. ! USE 1978 1983 PRINT pop_difference female_proportion ! ! Now save the items in the databank. Note, however, if you try to ! KEEP a data item that currently exists in the databank, you get ! an error message and the new data are not saved. ! KEEP pop_difference female_proportion ! ! Instead, you must REPLACE the data item, either with the REPLACE ! command, e.g. ! REPLACE pop_difference ! ! or KEEP the item with the ON REPLACE option enabled. ! ON REPLACE KEEP female_proportion OFF REPLACE ! ! Everything should now be set in our databank. ! CONTENTS ! ! To make sure that the two revised variables were replaced correctly, ! first clear the workspace and then print the items. ! FORGET pop_difference female_proportion SYMBOLS PRINT pop_difference female_proportion ! ! The only thing left to show is how to delete your databanks from ! within SORITEC Sampler. This is done with the PURGE command. After ! the two databanks we created are purged, this demonstration is over. ! PURGE 'a:popdata' PURGE 'a:ecdata' cls QUIT