How to create dictionary file from SEER that Stata can read?

[Arnaud Lesegretain]

Hi

I am very new to this, so hope the question is making sense.
From SEER, i was able to export a text file with data and a dictionary file. This dictionary file is .dic file and state does not recognize it. It seems like stata needs a .dct file.
Has anyone experience in either converting a .dic file into a .dct file or being able to directly get from SEER*stat the right file format?
best
Arnaud

[Mike Lacy]

I would guess that the subset of people on Stata who have used ".dic" files and SEER is not null, but is small. I think you'd improve your chances of getting help if you could post an example and/or description of the .dic dictionary file format. If .dic files are just text descriptions of text data layout, someone with a little programming skill in text processing, rather than specific experience, might solve your problem more quickly than waiting for an experienced person to take interest.

[Arnaud Lesegretain]

Hi Mike,
Many thanks.
have copy/pasted below an example of the .dic file from SEER:
best
Arnaud


[System]
Output filename=C:\Test_111821_Lymphoma.txt
Matrix filename=C:\Test_111821_Lymphoma.slm
Database name=Incidence - SEER Research Plus Data, 18 Registries, Nov 2020 Sub (2000-2018) - Linked To County Attributes - Total U.S., 1969-2019 Counties

[Session Options]
Type=Survival

[Export Options]
GZipped=false
Variable format=labels
File format=DOS/Windows
Field delimiter=tab
Missing character=space
Fields with delimiter in quotes=true
Remove thousands separators=true
Variable names included=false
Column Variables as Stats=false

[Variables]
Var1Name=ICD-O-3 Hist/behav, malignant
Var1DisplayType=Unformatted
Var2Name=Sex
Var2DisplayType=Unformatted
Var3Name=Year of diagnosis
Var3DisplayType=Unformatted
Var4Name=Race recode (W, B, AI, API)
Var4DisplayType=Unformatted
Var5Name=Origin recode NHIA (Hispanic, Non-Hisp)
Var5DisplayType=Unformatted
Var6Name=Chemotherapy recode (yes, no/unk)
Var6DisplayType=Unformatted
Var7Name=Marital status at diagnosis
Var7DisplayType=Unformatted
Var8Name=Vital status recode (study cutoff used)
Var8DisplayType=Unformatted

[Format=ICD-O-3 Hist/behav, malignant]
7996=Benign
7997=Borderline
7998=In situ
3=8000/3: Neoplasm, malignant
7=8001/3: Tumor cells, malignant
11=8002/3: Malignant tumor, small cell type
15=8003/3: Malignant tumor, giant cell type
.....

[Format=Sex]
1=Male
2=Female

[Format=Year of diagnosis]
200=2000
201=2001
202=2002
.......

[Mike Lacy]

From seeing just this piece of your .dic file, I have changed my mind. I now think that translating this file into a .dct file can reliably be done only by a person who has deeper knowledge of the .dic scheme. Although the "Field delimiter=tab" item makes me think that the data is just in a tab-delimited text file, the method of listing variables and so forth shown in your example is not transparent to me. Maybe someone else here will have a more helpful idea, or perhaps some more investigation on the SEER data site (about which I know nothing) would give more details, or, better yet, give an option to obtain the data and dictionary in a more well-known format.

[Arnaud Lesegretain]

Hello Mike, many thanks for taking the time. I will keep digging and post any solution I find !

[David Radwin]

Maybe there's a workaround in which you can export to a file format with labels that Stata can read, like SPSS or SAS, or can be converted to Stata format using a third-party program like Stat/Transfer?

. help import